Tokenomics Wit Paper | Difuon Fog—Dual Tokenomics: Redemptive Benefit Token Samuel M. Smith Ph.D.'

Original 2018/10/25, Version 1.3, 2018/11/29

Abstract— . A novel dual utility tokenomics model is described consisting of both a medium-of-exchange (Mok) token and a. store-of-value (SoV) token that together provide true separation of concerns between the need to support stable pricing, low Ficion, and High (Houghput ansacions and he necd fo apprecitve vaue fo evard vituous partcpation on e psiorm including investment. This approach avoids the inherent conflict between the two concerns that is the root cause of the failure of e Singh iy token mode ke osght s that opfonl iy is Sl . Onee sepated Fom he condants Sy low friction, and high throughput, the SoV token mechanism design can be fine-tuned to improving participation and value growth in the associated platform. A new complex SoV token is introduced called & redemptive benefit token that cmploys Sl complementay feutres ha i dowrmoand vlatily and driv »itious beavio and vale during dieret growih stages of the platform. We call these HoDling powers. This approach is general and can be beneficially adapted to many utility orem appiaton.

ndo Tems. Tkenomics, Dual Toke, Uity Token, Mechaniom Design, Siable, nvsibe, Separaton of Concerns

1. Uity ToKENOMICS

My Blockchain-based businss models a of some form of evo-sided markeplace o newrk. This s mors ormalycllod a platform business model [13; 32; 33; 44]. One the reasons for using a distributed ledger is to provide a trusted way for multiple penies 1 xchange vilu over a cvork by aacking e usouted ansnction n th eder Moreover th et he Sy Ty b cnbaneed i the govemmanee amd conol o e Aot odger doconalizd o Aiused aeros many pardn Obviously, a good way to facilitate and track the value exchanged in the transactions is o use a currency as the unit of account and e e of oxchanse. A 5004 medim of oxchanee b th ollowing faarcs: Iow Fiodon, widespred seces, High throughput, low latency, high velocity, low average settlement cost, and stability. Low friction means that the use of the medium O excange ighly comveent o (e paricpants. Widespron access mosms Tl Hkly partcpams have he porental 0 obtain the medium of exchange in order 10 participate. High throughput, low latency, and high velocity mean that turnover in s of th medim of cxchange i sapul 5o ool Wansectons st nor vpeded . the mechami (o sctlement of the ansacions Low avrage setlement con means et he o8 1 ove

shead coss on avcrase o win he modium of xehange fox v vamsacion re vy low rlatoe o he vlue of ot Gamcton. Finaly Sabity ey tht he chans i prc o praducs and services . he mdium oF xchange due 10 vy, apprciton, o depreciaion s i Fcam laie o he product life cycle. Fiat currencies like USD exhibit most of these characteristics, especially cash. The problem is that for a Renwork the vansctons s b lcckonic (g uch tht G i3 ne prachicl This ten rquies some clecronie payment channel which may introduce some latency to the transactions, and whose provider may charge relatively large settlement fees especially for micro- or nano-transactions. The requirement for an clectronic payment channel may often limit access by some pevicpants,especially the umbanked. Bt sl e bin s the ow i, igh sloety,and STty of it ke e primary candidate for an electronic medium of exchange.

The Feh hat the disbuted Iz can b sae 10 uppor ¢ cryplocumeney o s cryplographie oken has reted the pton where the medium of exchange is a crypto token instead of a fiat currency. The main attraction of a cypto-token as a medium of Cxchange 5 he porcatal 1 e loar ey and lowe selmen aoss v 1o comentons lckorc il payment Chammts. Anothratsacion has bt potental 1 ave pacudonymous ransactions e csh bt secson. & popula e o crypto token is a utilty or usage or network-medium-of-exchange token [8; 41; 52; 68; 72]. The term utility is used because the okt deivs vl rom s ity . he uncioning of e esoiated nctwodk o platfomn. Typically h pemary foncion o the utility token is as a medium of exchange and unit of account for transactions on the network. Another function of utility Lokens i o nectiviz the prticipation i . nctwrk 1 Suppost the it ldger nd olber related funckionalsy. This 1 done by awarding tokens to participants who perform these functions.

1. SINGLE UTiLITY TOKENOMICS

But what has really accelerated the adoption of utility tokens is as an alternative means for raising non-equity-dilutive capital inordr o fund the Srctio of e ssochied anline mrkeiplae (1114 411 This s done trough s src of ke genertng events, where Lokens are sold to investors that are then traded on exchanges. This type of raise is typically called an 1CO (Initial Coim Offrng. So ot combined 20174 2015 st 2000 sups ave esed over 13 Bilion dolls s 1C0s 301, In

Tokenomics White Paper 2 order for a utility token to be attractive to investors it must have an appreciative quality, meaning that the investors must be able {0 buy low and sell high. The hypothesis is that a large appreciative driver of price for the token can be created by limiting the supply and making network function highly dependent on the use of the token [20; 41; 72]. As the value of transactions on the network grows demand for transactions will drive demand for tokens which given the limited supply and functional dependency will drve the price upwards. The token then scts s sore of valu tha i appreciating over 7 9] The simple approch to forcing functional dependency is to have a single utility token for all functions of the network. This also simplifies the software ‘needed 10 support the network function relative 1o the token. This is the approach that the majority of the carly and existing token generating events have taken. A source of problem i the thought that ane should add as many sources of value o th token 5 possible in order to enhance the market value of the token [72].

“There are two scrious flave in this hypothesis. The frst s that mrcly lmiting the supply and functional dependency are ot sufficient to drive the price upwards, the velocity of tumover of the token must also be kept low [4; 12; 15; 68]. That is, relative 10 the latency of transactions, the availability of token reuse for subsequent transactions must be kept low. This is in direct conflict with the desiabilty of using a cypto-token a5 a ow ficion, o latency medium of exchangs. The second flaw is that appreciative price of the token is in conflict with its use as a stable medium of exchange. As the token appreciates in price, the price of producs and services on the network also increase making the network as a whole less atractve to alternative markets that provide compeiive products and services{12; 15; 40]. This becomes a disincentive for participants o use the network to exchange products and services. Only products whose demand is highly elastic relative to the price will be attractive on a platform shere the pricing is going up fast elaive o the product ifecycle and opportunity cost elsewhers. Together the low availability of tokens and the appreciative pricing are strong counter drivers to product demand. These means that the network Value never grows or grones so lowly tht the intrinsic demand for the token as @ medium of exchange never gros or grows slowly thu

s muliying the appreciative driver that was the morivation for using a iy token in the irst place as a means of vaising capital. I other words, it s not practical to use a single crypto token as both @ store-of-value (investable, appreciaive) and as 2 mediun-of-exchange (ow ficton, sable), Ancther way of looking at this is with Greshans v, hat sate, "bad maney drives out good" [10]. What this means is that if two tokens are in circulation with the same face value, the more valuable one will disappear from crculation, .. it will top being used as a medium-of-exchange. A good medium-of-exchange must be justa litle bit bad s0 1o one prefers holding onto it over exchanging it.

"The resut i tha the most of single utily tokens have fuled already or are predicted o fal in the marketplace [23; 36; 38 70]. The usual lfecycle of a single wtlty token follows: The iniial sale price to nvestors is usualy discounted from what the expected price willbe once the network is unctioning and is approaching its market potental fo transacton volume and value., Prior to netork going online, the tokens are traded on an exchange and the price is purely speculative. Hype and promion of the potenial value of the network (ump) incentivizes secondary bugers to acquie thetoken at a igher price, thus aowing the iniia investors to liquidate their tokens at profit. In most cases, the liquidation floods the market with more tokens than the demand can absorb and the price drops, thus incentivizing investors to liquidate even faster (durp),thus rving the price t0 a relative zero [19: 311. This is classc cxample o the greate fool theory in sction [25; 26]. The price charts for the vast mujority ofsingl utlty token raises have followed ths patern 47 Too examples are shown below [17]

EOS Charts P

‘ Natwork Live. f I j ’:N' b i H H il H E M y FW‘\' J Ky A - if N VAR v oo Pl a

fa X MJ\ i s |.\|\I\J|m|\“.‘ It e

Fi 1. FOS Pricchrt

Tokenomies Wit aper s

Tezos Charts LnearScale Log scaie 17 =

Luton .wg-‘ ey H i il e bkl b " s ! [ [ —————

A small prcentage of networks funded tis way have gone online nd are selig products and services. Tn many of hese cass, however, e he velocity of thefoken s high cnough thtfhe netwark s functional ithou (he Gken price spprciaing back 1o anywhere near i pretaunch Ievels or he network pricng has spprcited making the prodacs and seices incompetive. Tn o ncble case, QuaniSamp, the network added o Fat payment chamnel in Grde 10 esors price Compentivences (6], TH additon emoved ot of the vlue o th token (e ntwork .10 ongee funcionally dependent on i mon-rvial way) Pio o s change he Quansiam token had boen genraly apreciatng. ARerwards the pice declned never to recover

ntstamp Charts Unearsaale Log scale 10 = Quantstamp Chart Begin Acoepting Fiat (USD} e s [ ———————— Fie . QuantSang (80 i Chan “The failed actual performance of dozens of networks over the last few month due to these characteristics is a strong refutation.

Tokenomics White Paper 4 of the single utiliy token hypothesis [35]. 3. Mueri-UtiLity TOKENOMICS

A more complex nuanced approach is to use two or more tokens for the function of a given network. Given the fuilure of single uility tokenomics, a motivation for a dual or multi-token approach is to support the separation of concerns between the use ofa token as a medium-of-exchange (Mok) (non-appreciative, stable) and a store-of-value (SoV) (appreciative, investible). Somewhat confusing is the fact that some networks like Ethereum (Cther and Gas) and VeChain (Vet and Thor) use dual uility tokens to separate other concems such as a floating price relative to Fther in Gas for the spo pricing of computation on the network independent of the pricing of products and services provided by the network [S1; 62; 67]. Such is an exanuple of dul utilty tokens for two different network functions. Also confusing are dual tokens of different types. The first token is a utlity oken, which is used as a medium-of-exchange on the platform. The other token could be of a different type such as a security token or an asset backed token [1; 27]. A security token has rights to equity or share of profis and is not part of network function but appreciates based on the vale of the equity andor profits generated by the nefwork [65: 66]. The problem with security tokens is the lack of viable security token exchanges [5]. An asset backed token derives its value from an associated asset that has intrinsic appreciative value based on property (real or intellectual) or a scarce commodity [3]. OF note is the distinction between a tility token thatis sold as a security and a securty token. Under US securites law the Howie test determines whether Something is a security or not. Simply if token is sold that does not have true consumptive value o the buyer does not have reasonable cansumptive intent then that token could be a security. Before network or platform launches a medium-of-exchange token has no consumptive value and is therefore its sale s likely to be considered that same as the issuance of a security [34; 42; 71]. Once the netwrk launches then sale of the token might not longer be considered a security issuance.

o that end, a better line of inquiry is: *Can a digital asset that was originally offered in a securities offering ever be later sold in a manner that does not constitute an offering of a security?” In cases where the digital asset represents a et of righs that gives the ho

lder a financial inferest in an enterprise, the answer is likely “no.” In these cases, calling the transaction an inital coin offering. or “ICO.” or a sale of a “ioken,” will not take it out of the purview of the USS. securities laws. But what about cases where there is no longer any ceniral enterprise being invested in or where the digital asset is sold only 10 be used o purchase a good or service available through the network on which it was created? I helieve in these cases the answer is a qualified “yes. "[71]

Herein we focus on multi-utliy tokenomics for the purpose of separating the two concems of medium-of-exchange and store- of-value. The core thesis is that an appropriatly designed dual utily token model allows one of the tokens to be a storc-of-value utilty token which provides liquidiy to investors when traded on available exchanges (unlike security tokens) and may appreciate s a function of the valuc of the associated network while the other token acts as 4 low friction high velocity medium- of-cxchange thus avoiding the drawbacks of a single utility token model. The key insight s that optional utiityis still tility. The network can function as a successful compefitive business consummating transactions merely with a medium.-of-exchange utility token without requiring a store-of-valuc uility token. However, the storc-of-value token, although optional, can provide additional value to the participants in the network and to the function of the nefuwork that may be appreciative in nature [38] Tndced. the flawed hypothesis that required uiiliy i cssctial forces the inherent conflct between appreciation, pricing clasticity, and velocity that has resulted in the almost universal failure of single utility token models for anything other than a5 pump-and- dump speculative vehicles,

‘The medium-of-cxchange token could be fiat. ndeed, given the low friction, relatively high stabilty, high availability and low latency of fiat, it is difficult to argue that the medium-of-exchange should be anything ofher than fiat. Currently most non-fiat tokens have super high friction to on-boarding participants and converting their fiat to non-fiat uility tokens. This friction has become even higher in the USA, because the IRS classifies non-fiat crypto token exchanges as asset transactions subject to capital gains taxcs. The biggest molivation for using & non-fiat token would be lower lateney and lower scitlement costs than existing clectronic fiat payment channels. B

ut in cither case, because the world runs on fiat, having 2 transperent fiat payment channel to access non-fiat medium-of-cxchange token is essential to universal adoption of the network. One way to achieve both stability and lower settlement costs is to peg the ulilty token to fiat for stability, but batch the conversion to and from fat to reduce scttlement costs. The simplest pe is an cphemeral two-way one-to-onc peg with a liquidity pool in fiat. Ephemeral means the token is minted and burmed on demand. A fial-pegged stable ephemral uility loken is nol appreciating or depreciating by design o it is mot investable or tradab, but may only be used to acquired products and scrvices on a platform. These arc good Teatures from a regulatory perspective. This type of ephemeral peg is in contradistinction (o non-cphemeral but sable crypto- currencies that are meant (0 be traded [18].

An cxample of utlit token medium-of-exchange with a fiat payment channel is the Stcllar nctwork, whercin fiat anchors provide credit from fiat accounts for end-to-cnd fa-to-fiat transactions that use the utilty token (Lumen) as an ephemeral ‘medium-of-cxchange [56]. The fiat anchors provide fiat liquidity pools to back the credit they extend for the duration of the transaction seutlement time (< 59). The result is that Stellar act as & transpareat, low [iction, low latency, fat payment channel with low average scitlement costs. The Stronghold network on op of Stellar is providing a USD peged anchor [10. Utility tokens used as & medium-of-exchange would want o exhibit similar features.

‘There are other more complex ways to implement a stable ulility token such as using an algorithun for fine wning dynamic supply that may make the token even more stablc than fiat, However, for many two-sided network business modls, fiat is sulficiently stable such that a simple peg i al that is needed if not fat isell

Tokenomics White Paper s Given that low friction, high velocity, stable medium-of-exchange token (which may be far) is used fo consummate transactions and pay partcipanis for supporting the netvork, an optional companion storc-of-value uility token can be added without conflict with the function of the medium-of-exchange token. Primarily the optional uility comes from using the store-of- value token 1o promote and reward virtuous partcipation in the network. Such token may provide value 4s & marketing (ool to incentivize adoption, shape behavior, and reward participant-sourced improvements t the nctwork. These arc ll good drivers of the positive two-sided network effects that enhance value of the network. These drive the network dynamics with positive feedback loops. The goal is that the platform via ts mechanismm design and associated slgorithms may automatically result in appreciation of the SoV oken as the network grows. 3.1. Redemptive Benefit Token

We have designed a new kind of optional store-of-value utility token as a companion to a medium-of-exchange in a dual token model. We call this new store-of-value token a Redemptive Benefit Token (RBT). Mechanism design is a temm used o describe the process of coming up with the functions of 4 token in a system thal drive il (0 allin ceriain behavioral characteristcs. The redempive benefit token s based on a mechanism design with murually reinforcing feedback loops that incentivize desirable behavior and de-incentivize undesirable behavior in the fokenomics. s primary goal s 1o incentivize growth and good behavior of participants in the network while limiting downward volatilty and allowing for upward appreciation in the price of the foken. The token benefis both its holders and the network. In homage to the internet meme resulting from the drunken misspelling of hold as "hod!” 1o deseribe the practice of holding onlo store-of-valu tokens, which serendipitously s also an acronym for "Hold On for Deal Life" (HODL), we have filled these mechanisms the “HODLing powers” of the redemptive benefi token [24; 28; 61].

3.1.1. Hodling Powers of the Redempiive Benefit Token

The redemptive benefit token (RBT) approach to providing 2 store-of-value (SoV) token that works with 2 companion medium-of-exchange (MoF) token combines the beneficial properties of git card, loyalty, rewards, and staking programs to incentivize acquisition and holding of the SoV’ token i the following ways:

(1) Redemption: Like 2 gift card

SoV tokens can be redeemed at the rate specified by the MoF. tokens for items on the platform’s products and scrvices menu. The men s two-sided with one side priced in the SoV token and the other side in 2 MoV token which s pegged to fiat. This provides a composite floor value of the SaV in fit that is equal to the consumptive weighted average of the fial valuc of the lcms in the menu. This is designed to minimize downward volatility and help ensure purchaser's of SoV tokens who purchase at a discount realize positive retums and also virtuously ensures that ltcr purchasers of SoV tokens arc protected from downward volatility. The redemptive valuc is effectively a worst.case value for the SoV token once the network is liv.

(@) Promotion: Like the points in loyalty program, SoV tokens may be redeemed for MoE Tokens for products and services at promotional prices provided only to members of the program. This allows conversion at beter than retil rates and allows for anti-Moore’s Luw appreciation as service classes through the Menu are upgraded over me in response to technological advancements. This drives carly adoption by users of the Project and cnsbles repeated adoption accelcration as new products and services are introduced to the neork through the menu at promotiona pricing.

(3) Rewards: Like a rewards membership, holders of the SoV may cam MoE tokens as of function of both ther holdings of S0V and their spending in MoE on the platform. These camed Mo Tokens may then be redeemed in exchange for Fiat at e rate specified in the Menu or used 10 purchase additional products or services on the platform. The carning ate of Mol: per expenditure of Mol is an inverse nonlincar function of the number of SoV they hold. For example, a holder of 10,000 SV would earn MoV ata higher rate (but less than 10x) for the same expenditure of Mok than a holder of 1,000 SoV. The cxact nonlincar function and parametcrs is tuncd {0 the market dynarmics. An cxample of a nonlincar function is given in scction 4.2 below. This incentivizes large users of the platform to hold increasingly large numbers of SoV in order to obtain lrger rewards in MoE, The amount of SoV held incrcases a5 1 function of the markel cap of the network in services provided. Holding of the SoV token can be further incenivized by increasing the discount as a function of the time the SoV token is held

(4) Membership: Like membership in a club, partcipation in several activities within the network require t

he stking (holding) of SoV. These include, cither dircetly or as a delegate, the hosting of support nodes, the voting on the use of community tokens for bounties, or the voting on features and upgrades 10 the nefwork protocol. This staking rewards good participant behavior by providing the opportunity (o share in the revenue generated from fees associated with supporting the network and dis-incentivizes bad behavior as participants wonld lose their stake as a result. The participation strengih is an inverse nonlinear function of the amount staked. This also reatly incentivizes the holding of large numbers of SoV that increases as # lunction of the size of the network infrastructure.

Given that the supply of SoV is limited, then the holding drivers above would increase demand for SoV’ as the transactional

value of products and services exchanged on the network increases, threby providing an appreciative driver o its rice.

The Redemption HODLing Power above is designed to limit downward price pressure when the network s small whereas HODLing Powers 2, 3 and 4 above are designed to provide upward price pressure for SoV tokens well above their redemption value as technology improves, usage expands, and particpation in the platform increases. The intent of this design i to produce a price curve with beneficial characterstics. The Membership HODLing power uses staking of oV’ 10 ensure good behavior. It

Tokenomics White Paper 6 can be thought of a potentially refundable license fee or deposit that enables an entity (0 host a platform node and receive transaction fecs for that hosting. Should the catty misbckave then they lose both their membership and their staked SoV fec! deposit. A notional example of the priing behavior due Lo the combined eflect of the four HODLing powers over time is shown in the following diagram. Price ! Pre Launch | Redemptive Promotional Rewards Membership Phase ! Phase Phase Phase Phase x-10x 1x-5x >5x 10x - 100x Network Launch Discounted | Redemption Value Price . Time Fi, 4. Notonal phased prce versus time for So¥ oken with HoDlng powers.

Discounted pricing of the SoV may be viewed us @ type of warrant on fulure margins of the products and services. The number of discounted SoV tokens must therefore be limited and only be available prior o network launch during a pre-sale with a cap on ‘money aised. To limit the rate of redemption of discounted SoV tokens and further incentivize holding, time locks andor time Variable redemption rates can be employed. A time lock can freezes or lacks a token so that it cannot be redeemed uniil the lock expites. A time variable redempiive rate increases the discount as a Aunction of time up 0 some maximum, This allows for immediate redemption but rewards the holder for waiting longer to redeem.

“Thete are two main paths to obtaining Mof: Tokens in order 10 purchase products and services off the menu. The first path is by purchasing SoV tokens and then redecming them for Mol: Tokens that may then be used to purchase products and services from the Menu. The second path s by purchasing MoF: Tokens for fat directly. The fist path may reward investors and other platform participants. The second path may enable wider adoption through more conventional retal fat purchasing. Products and Services are priced in Mok Tokens and the Mo price is pegged to fiat. This sabilizes the price of the Mo foken so that it is less susceptible 10 the pricing volatility which could hurt the competitiveness of the platform. Indeed, the Mot: token could be fiat, A 00 peg is USD because of its ubiquity s reserve cumrency but may be changed o one or more llernative fiat currencies on a eographic basis or 10 a baske of fiat currencies o some other pegging mechanism, The goal of this approach is a more stable price for products and services reltive 10 fat that i decoupled from the price dynammics of the SoV token.

Once th

e network is launched, the platform may offer additional retail SoV’ (at their redemption price or higher) in limited amounts 1o ensure liquidity pay usage based camed zewards, and as bounties for team and community effors (© build and enhance the platform.

‘An alternative approach is to not enable paths through the two-sided menu from the SoV: side o the Mol:/fiatside. One reason 0 do this would be to remove the fiat liquidity pool requirements for issucd oV’ tokens for iems on the menu. This means that producers any item on the menu would be willing to accept payment in SoV and not MoF:/Fiat fo that item,. If not then the menu item is not aceessible from the So’ side. Likewise o meau item could only be accessed from the Mok: side and the producer would not aceept SoV’ tokens. “This could enable the platform time to build up liquidity. This would also simplify the demands on producers who wish to put their own items on the menu. The problem however is that it imis the accessibiliy of partcipants 0 products and services on the mem. In other words, some items can only be purchased and paid with MoF:fiat and some items can only be purchased and paid for with So.

Tokenomics Whi Paper 7 In any event accounts are differentiated by the path through the menu for sake of refunds. The path out is always the reverse of the path in from the standpoint of refunds. This means that transactions maintain the path through the menu unil final settlement. For example if & purchaser has 60 days for a refundthen the ransaction pat must b retained for 6 days o enable a e o retrace the path in. If & refund does not retrace the path in then participants might try to arbitrage different effective SoV veluaions fo dilfyent s it e, T ey dofot e prormotional valu of variable priing i the mean. This i ot a problem for MoE that was the result of a fiat only purchase but only Mok that came from an SoV redemption. SoV redemption can be thought of as creating an MoE voucher. The voucher persists until final settlement including any refund time periods. For example should someane redeem S0V in build but noF consume all the services immediately then some balance will be outstanding. 1f at some time later (within the refund) window, the purchaser desires a refund then the refund amount is the famount given by e original redeaption vouchie pats o SV 10 MoE. This moans that purchasess do ot havo an accouat with undifferentiated Mok, They have pools of Mok corresponding to the path into the purchases. These include a fiat pool (came in from fiat), and a separate pool for each menu item path they used to convert from SoV to MoE. Functionally SoV are Very much ke poins i loyaly program and menu ems represent specal certificates and vouchers tht result from redeeming points for services. The points stay points until they are consumed with finality, that is, a final un-refundable purchase. 3.2, Use Cases

Considr for exampl a consumer fhat wishs o purchase productsor computesevice isted i e men:

• The consumer i not imerested in paricipting in he plaform other han merely buying compue services at he prices affae i e e, Tho sharing; ccononzy s, of e Rctwork sbould maks oz of foes cevioo sorapeiively piond. Using the fiat purchase paih fhe consumer <an e buy. those services wsing & similar payment channe fo cising aliceaatives, T a payspeat chanael tweaspascily couverts S {0 Mo ok hich ar used 48 median of cnebaags and unit of sccount o the et This payment channcl seeksfo minimizo the friction o he consumer

« Alternatively, the consumer may prefer to take advantage of the promotional pricing that may be available via an SoV purchase fom tho platiorts

irotery o the consuaue 18y fasiad b0 able o by 4n SoV tokca Gom st clo a & discount to its redemption value for products or services on the menu. The menu is two-sided with each item priced in both. S0 std MoV (okena. Some ieas o the sacna gt be of ore use 10 given consunnes based ot the comsumer’s ed and thereby might offer better value when purchased via the SoV channel rather than for fiat. A consumer, may be able to create better overall value for themselves. rather than by using fiat payment channel.

Consider anather example where & consumer bought SoV at discount r even atrtail face value but held ontathe untlthe network has grown and the rwand or staking detmand for the SoV hasrised theprie of the SoV token on an exiertal sxchange o above the redemption value of tose SaV' (okens via the Mema. Farver, suppose the consumer nor desires 1o purehase products or servies fom the platfo. Given the S0V value o an extemal exchange is sbove the redemption value from the hetwork fo fhe SoV held by the consumer he consumer may nsiead s {0 obain a greater mamber of products ar servces for those SoV. than thei face value by first selling some SoV on an exchang for fiatand then usethe fia payment channel (o buy Mo tokens to obtain the products or services from the network.

Suppose insead that a age enterprise consumer of ervies rom he placform eshes fo bian thos services a a discoun o el retal price. The SoV rewands program provides Mok tokens as & reward fo the purchase of services as function of the umber of S0V iokens held at e ime of such purchase. This ofctvely disouns the pice of toe sevices for such a consumer. The enterprise consumer would then buy and hold SoV tokens t enable the rewards of MoE tokens every.time the consumer purchased Mol: tokens and hence services using the fiat payment channel. Notable is that the consumer does not redeem SoV okens in thiscase but nsicad hold them in orcer o get Fewards of more Mo tokens fo te purchases of services through the Project.

Finally, suppose s aricipant an he nework wishesfo have a new fsture added o he nework and wuld Ik fo vot for hat feature. In this cas the paricipant buys and holds SaV tha it stzkes SoV i order tht the participantcan vote on the new feaure o be paid orwith SV tokcns fom the communiy allocation 3,13, Simpifed Examples

The redempiion and promotion HODLing powers rly o tworsided menu. A simplifed cxample i shown in the ollowing iagram. T the cxamples,the cutency symb for MOE.Fiat s S a

nd th currency symbol for So s

Sov [ Menu | MoE MoE - Fiat

| Products | § Liquidity Pool

[ so | Sov [3] P T3] Reserve Pool [« m T4] Fig 5 Tl M Cner ol s roduc s o pe o, The sl i the e i SoV The i ol s e i

Tokenomics Whit Paper 5 Mo Al hown i he s po e ecad S s iy ol of ot ke Mo o el In the following diagram is shown an example fiat transaction using the MoE side of the menu. Sov [_Menu [ moE ME |y [ MoE-Fi | 4

_[Podcs | S 2 A s¢ [W¥ vy ool

e ] [so] S s ROsam [ s2 ] N I N AW e s Bob Sam Mok Fie 6. Txsnple Mol oo ol coded s oot s e o f s - 1 Al oo e by poe 2 o e i LS ey 4 ot B2 B ol ks b I il o, o s pot s o 4 o e kom B3 s 5 Vo pe i e B s o b o, (e vl 2 Al s i St opai o o spprin e s 5 5 ol s o o S . oty bt 5 s et ot 1o el o R ol T e 5 S ot o 3 . i ol 1 the following diagram i show n exmpl redempion runsaction using e S0V sid of the men Sov [_Menu_wioE (o] MoE— Fiat 1 # | Products | $ 2 53 | Liguidity Pool = [ ] Coe ] “ s st KN ER o jred N Vs s [ ™ [s1] Vi 7. P 5o o i, Cloe o g how s e shcndfch . S S =5, B = 51, Carol = e b 4o o e b et i e S S o e 4 S b ey B pltiohietodibohe elaa b Ko e it P it ety b g e ot ek 15 e et el B s T S B o, i e o e st o e S e et o e S S B i e e o 3 ey o g e et s o ok B 5 o e s o e s s ot e it SO AW oo s oS o e e et o s S e e o e o o e g o o 1 1 the fllowing diagram i show an cxample promotona redempion tansaction using he SoV sid of he e

Tokenomics Wit Paper 5 1 SoV | Menu [ MoE % MoE - Fiat L $ 2 Liguicty Pool ar e T s/Tse] =N\ =1 MoE [ = =] [se] ’ e e B o T 4 e e o s et o oy 514 Summary

The gol of the dual oken mdel is o provid mre sophisicated (okenomics behavior ha rewards and incentivizes particpans o the nenwork gows e aso provid for competitve overal pricng.

o reter, o Mo, 1okins s pesee 10 o a B s xchange . Associated withthe it paymentchanel wil e hauidy pos of fat When consammrsofsorvies oginalty buy MoF tokens wing th it channeh o or sl of he it ned o prchase thoe MO tokenswil g0 o 1 gy ool The asocited Mo tckensar then ated 10 by services The o providers oot MOE o exchange forprovai serices. From e 10 . th s provides may oo choos fo echange hts MaF back tvangh e T iyt channe andreceive ik o e iy pool The moraton for wing SEoF tokensas & mdiam ofoxchan mtoudof Fat s tha thstlomnt costs nd ateney of h it payment hanncl ity be 0 igh for th mame ansacions neaded 1o suppor seviecs. The ot M. comirsions andtoe MoP 0.t comerions e ched t reduce coss and mprove rotghpas Over e, he quidy poo may have an oxces tht 1 2 el o h sro it e st seices oo 1 s and wha S provdte ar g THE oxcon co tho e used o thr Panporcs whhout negaivls afscang MO iy

T tokenomic of Sov Tokens 15 wholly disint The platorm il rdocn S0V with MoF using the men valus of the ssocatod sovion ot the fime of purchse (whichca b +41ad o he comeapanding peed it vau). Ths effetive e esed oxchamge st may b Sgnihcanty e rom he oaing cxchange e bemeen SoV. and i on 3 o iy Crehange, The ssociatod MoF e the s 1 by scvics. Th scvi provider rceive MoE i xchange 1o proviing emices. From e 1o e, he s rovidrs iy h choose o oxchange it MoF bnck i it hrough he ot paymens Chamne and rccive Bt o he iy oo, The Ty poo will ncd ancugh ences 10 cover the vl comersion o HoE back i i that wer st sourced a5 SoV. This cxeces may come from funds s on he il af SoV o from he acoumted gros masgin o 1 punchasedsevics There nay b some hauiy sk shoul e enoun o SoV b ek o iscoun grencs o e pivalnt oo i of . esouted seices an e ment. Th disoumted SoV iy ity ‘may be covered by the accumulated excess in the liquidity pool resulting from the gross margin on services bought by others at a. fes dcomed pace.

At some e e T, once th ltfom i v nd experincing sy v growh and profits and with approprite regulatory approval, the platform may choose Lo also provide more convenient exchange paths for participants at the prevailing Cheral oxchame ot 1o ety oxchinge MoE

for SoV. o S0V o i, or ik o SoV. The reason I proviin Such exchange paths would be o reduce the friction experienced by participants in order to increase the rate of adoption and the size S posiie ncwesk sl ecdback Inlly, Bower,por (s etwork lnunch he ety rovidd pat of oxcrans will b (1 SOt o MoF 10 i and (2 i o MoF 10 it T each eaes e it comserions il b g h i pymen channe and liquidity pool. Direct conversion of MoE 1o SoV, SoV to fiat, and fiat to SoV, and will not be initially supported by the platform and st perlormd on . hd-paty cschange, f .

1. BALANCING PARTICIPANTS INTEREST AND INFLUENCE

The main pupose ofthe SoV {oken i o incentvize benefcal paricipton on he laform, The oken i usd t posively reward good behavior and negatively reward bad behavior. We call the anticipated rewards or return of a participant their interest o meres. I desentrlned e mlipl e <ol he contol that some. prtipas ight have ver ome compones o the network or the behavior of other participants, their influence. Its usually bad for the platform when the interests and influence

Tokenomics White Paper 10 of some partiipants are such that they can use their influence to enhance ther interest unfaisly. Consequently one purpose of ‘mechanism design s to build checks and balances on perticipants interests and influcnce. For cxample, with the first two "HODLing powers, redemption and promotion, S0V holders do not have much influence and their interests are simple. They ‘mercly may choose to make arms lenglh exchanges of okens for products or services. The exception is hat when the platform is small, a holder of & large number of SoV tokens could pose a thrcat by redecming to many tokens rcltive Lo the platform's capcity or liquidity. As previously mentioned, one way to balance this influence is 1o lock the tokens for some minimum time period o to have an increasing redemplion ratc that penalizes/rewards carlylate redemption. The last two HODLing powers, rewards and staking, on the other hand have much higher potential for beneficial or detrimental effects on the platform.

4.1. Privacy and Consistent Pseudonymity

One complicating factor in online platforms is privacy. Because all transactions are performed electronically over a network it is easier for third partis 1o monitor and correate transaction activity thereby reducing the privacy of the primary partes and putting them at rsk of exploit by a third party. Exploits range from nuisances like spam more serious harms like slander and theft. To better describe the associated issues we first define some terms. We use a simple definition of identiy, that s, an identity consists of one or more identifiers and associated attributes [53-55; ST 60). An eniity may be any person, group, organization, computing device or soffware program. An entity may control or be associated with many ideniites. The set of all identites for an entity is its identit graph. Typically governments allow only one legal identity for an entity. Use of a legal identity puts the entit atsisk to legal recourse for il associated transactions. Use of the legal identity also puts tho catty at risk o exploits by malicious third parties. This is a two edged sword. In a transaction, knowledge of one party’s legal identity by another imposes a check on the one party's behavior due (o the risk of legal recourse for misbehavior. The use of the legal identity increases trust and facilitates the iransaction. This is an incentive to use a legal identit. Indeed the POA nefwork uses the risk of loss of reputation associated with v

erifiable legal identity as type of bond 10 ensure good behavior [45; 49]. Unfortunately, he risk of oss due to third party exploit may far outweigh the benfits of using a lega identity on a given onlin platform.

An alternative s 10 use & non-legal but consistent pscudonymous identity for online ransactions. A consistent pseudonymous identity s based on a verifiable unique pseudonymons identifier. Typically this is a type of cryptonym based on a cryptographic publiciprivate key pair The emerging decentralized identifer (DID) standard is the best example of this type of identifier [21; 57). Doens of arganizations are now either building portable identity systems or using such systems based on DIDs [2; 22; 37: 43; 59]. Consistency means the ety uses the same identifier for all ransactions on the platform and all partes can verify the identifer is being used by the entity with the associated private key. Consistency allows a reputation to be associated with the identity which can build trust. Consistency also allows the platform o reward behavior of a given identity thereby also building tsust. Pscudonymity preserves some privacy by maling it more diflicult for third partis 1o correlat 10 the entity's egal identity or other off-platform identites. Privacy may further be protected using 7ero-knowledge proofs (ZKPs) of credentials atiesting fo an enity’s identity [49]. The Sovrin network is designed 1o betier manage privacy loss using zero-knowledge proofs [60]. This comes at the cost of increased complexity and computation. The drawback of allowing pseudonymous identites on a platform is that entities may create multiple disposable or "cheap” identites that enable them 1o engage in bad behavior wih litle consequence to the entity. One form of bad behavior is a Sybil aftack where the entty creates sufficient numbers of pseudonymaus partcipants to unfairly distortthe entty’ influence o to overwhelm the network 64]

One appronch to this problem is to foree disclosure 1o a unique legal ideniity. This s the approach that banks must use in order to comply with legal regulations to impede criminal activity. This approach is called K'YC/AML (Know Your Customer/Anti- Money Loundering) [16]. The problemn is that legal entities may be organizations such as corporations that may be controlled by other entis. Setting up multple legal organizations is more costly than creating pseudonymous identifiers but may sill be relatively inexpensive. Because natural persons are uniq

ue, the ultimate nexus of interest andior control will reside in a set of natural persons that are the ultimate beneficiaries of the organizational enites. Consequently KYCAML requires disclosure of the natural persons that are the ulfimate beneficial owners (UBOS) of organizational entities[29]. This may be a very high frction high cost way of preventing the use of multple cheap pseudonymous identities for the same cnity (even via private disclosure via 7ero-knowledge proofs). For many platform business models it may not be practical or economically feasible to use this approach, that s, proving traceability 10 the underlying unique legal identites. Instead a lower friction approach (o mitigating the visks of muliple cheap pseudonymous identiies may be to use economic incenives that reward the concentration of activty into consistent pseudonymous idenites. For other platforms, especially where higher value participation oceurs, it might be appropriate (0 use & mix of cconomic incentives and legal idenity proving 1o mitigate the risks of both privacy loss and bad behavior. The next section describes horw to build in economic incentives for concentrated activiy:

4.2. Concentrated Interest Via Nonlincar Rewards

One way to penalize the use of multipl pscudonymous identitics by the same cniity on a platform s o reward the concentration of acivity (interest) into one consistent identity. A consistent identity means that al the activty i atributed to a verifiable identifier or a se of linked verifiable identifers. Attribution of activity to consistent identity is a very basic form of reputation ‘Ultimately, what any party uses to decide whether or not to cngage in an online transaction with another party i the reputation of the other party [53]. Suppose an enity is considering splittng its activity on a platform into two or mor partcipant identitcs. If the the sum fota interest of the spit prticipantsis less/greater than wha it would be f the activiy were concentrated into one participant dentity then the enty has an disincentive/incentive to spit. 1 on the other hand the the surm toal interest of the splic participants s the same as that of a single concentrated participant identity then spliting makes no difference in the revard

Tokenomics White Paper i Other factors, such as a desire 1o engage in untraceable bad behavior, would then govern the decision to split or not split. This latter case is true when a reward function satisfies the superposition property or principle [48; 63]. In other words, the sum total interest s the same for ether split activiy or non-split acivit. When the reward function saisies the superpositin property, there is no penalty with respect to the reward function should an entit choose to enable it own bad behavior by spliting its activity across cheap pseudonymous identities. Consequently if some other fictor provides an incentive (o split then the reward system does nothing to counter that other incentive. The superposition property or principle provides a departure point for analyzing reward functions with regards to participant splittng, The definition of the superposition princile folloes: Given a system reward function R with input x and output y at time . Let »(k) = RIx (k)] a 2, (k)= Rlx,(k)] @ x,(k) = ax, (k) +a,x,(k), @ where @, and @ are arbitrary scaling constants. Then we have, 25(k)= RLx, (k)] = Rla,x (k) +ayx, (k)] @ The reward functon R stisfics the superpositon principl i the output also stisfics, 7(k) = a,y,(k)+a,y,(k) = a RLx, ()] + a,Rlx, (k)] ©®) This result leads to the following defintion ofa inear system. A reward system is linear if and only i, Rlaux,(k)+a,x,(k)]= a,RLx, (k)] + a,Rlx, (k)] © and, R[0]=0 Ul Using induction a system with this lincariy property can be cxtended to any lincar weighted combination of inputs where e the weighs. In general,u finear roward system satsfis th following, . . P RY ax®1=E ar =% a R 0. ® with, R[0]=0, © When equation Eq. (8) holds, that i, the reward systemis lincar, then spliting the sctivty betwen any number of participants provides the same total eweard as combining al the activity into one participant. Thus only when the reward system i non-lncar s there the potential to differenially eward or penslize concentzation of ctivily. Suppose for caample that the reward function of the form y=bx 10 where b i a constant. This is a common type of reward program where the partcipant receives a fixed percentage of their spending us "cash back”. This i lincar system. Suppose that the entity wishes to split tha activity among; two partcipant denttics under its control The tota reward is given by (k) below: FE =Y =X bk =bY, x,(k)=b(x,(k)+x,(k)) an Suppose an entity' total activty is $10,000 and splts it

with S8000 for one and $2000 for the other with b equal to 0.03. The individuals rewards arc 0.03(8000) = 240 and 0.03(2000) = 60 and the total, 300 is the same as 0.03(10000) = 300. With this type of eward function there s no disincentive to spliting. Suppose insicad that the Lollowing nonlinear ewrd function is uscd: y f y=] 02(r+(10000x) %) x<10000 P 04x x210000 The function in Eq. (12) is graphed below. The seward function is nonlincar up o 10,000 in sctivity and then is lincar above 10,000. Any split in activity below 10,000 results in less total reward to the entity. Once the entity's total activity reaches 20,000 it may split into two participants cach with 10,000 in activity with no loss in total reward. This function incentivizes a minimum level of activity and hence reputation for a consistent identity before a split. This penalizes cheap participation and rewards concentrated behavior up to a point. Many other similar nonlinear reward functions may be customized to a particular applicaton.

Tokenomics White Paper 2 Reward Function 0000 ss0.00 50000 5000 40000 000 » 30000 25000 20000 15000 10000 5000 0 1000 2000 300 4000 5000 6000 7000 8000 9000 10000 11000 12000 13000 14000 15000 Fig 9, Nonlinar Rward Funcion. Nolines below 10,00 sctiviy. Linarsbovs 10000 in iy The former cxample of & nonlinear reward fumction was dependent solely on partiipunt activity in terms of purchising products and services. This could be handled solely with the MoE token. An alternative would be to make the reward function also dependent on the number of SoV token held by a participant. The idea is to incentivize concentrated holding of SoV tokens by increasing the size of the revard in Mok fokens as a function of the number of SoV tokens held. This makes the reward function nenlinear with respect to participant activity in MoE as function of the SoV held. A simple example would be of the form, y=bwx ) where v s the amount of S0V’ tokens held, i the activity in Mo': tokens, b s a scaling constant, and y s the reward in MoF: tokens. An entity may now split both their store of SoV and their activity across participants. Splitting their store of SoV, however, rediuces the reward rae fora given amount of actviy. This penalizes splittng An example reward system of this type is given below: 001 T |x w<4000 y=1 (1000 . a4 04r w24000 where w is the amount of SoV’ tokens held, x i the activity in MoE tokens and.y is the reward in MoE tokens.In this case the reward function is nonlinear up until the amount of held SoV reaches 4000 and is linear above that. This rewards concentration. of behavior up to a point and penalizes cheap paricipation. I lso drives demand for SoV tokens, The cost of purchasing SoV tokens can be amortized over the expected increase in reward for a given level of participant activity. This allows mechanism design of the totalsupply of SoV that will be in demand as function of total activity on the plaform. Indeed this is the cssence of the third HODLing power. This HODLing power can be further cnlhanced by mking the reward not just & function of the amount of the SoV held but also the length of time that amount of SoV is held. This makes the SoV "sticky". ‘A more sophisticated approach could combine Eq. (12) and E. (14) into one revard system as follows: 001 ’ min{ | 1w [x,0.20e+(10000%) ) | w<4000,x <10000 y= 1000 as) 04x w24000,x 210000 Using the HOLDIing power of a rewards program based on SV tokens has the potential o both incentivize good behavior o

f partiipants nd good priiug behavior of the SoV token. 4.3. Concentrated Influence via Nonlinear Staking and Rewards

Tokenomics White Paper 13

‘The fourth HODLing power is via membership staking. By staking we mean that holders of the SoV token must put some of their SoV tokens at steke in order to jin the plaform and scrve in specific support oles. A stake i ke a refundable license foe that enables a participant to engage in a support role and get paid for that support. Like a deposit, the stake ensures good behvior because the participunt risks losing its stake should it misbehave. The main difrence between this fourth HODLing powr and the third i that the types or partiipation arc more involved than mercly buying or scling products or services Staking is used to modulate support and governance type participation on the platform. This includes delegation of stake to other partiipants s well us runaing nctwork nods. Staking is 1 means to manage participant influence on the plaform. Instcad of interest, we use the influence or the intensity of influence and the affiliated benefits as the return for putting SoV at stake. We can ‘model this elatonship s follows:

(k)= 1Iw(k)] a6 where z i the intensity of influence on the platform, w is the size of stake, & is the time, and / is the influence system function. The same superposition principle based linear vs. nonlinear relationships apply here as above but to the splitting not of activity but the spliting of stake. Recenly the concept of quadratc voting has gained attenton as 3 way of allowing incrcased intensity of influence in voting as an inverse quadratic function of the size of the stake or cost of a vote [39; 46: 50; 69]. We can represent the quadatic voting function as an influcnce function as follows:

2(k)=1Tw(k)]=(w)) an where 2 is the intenity of influence on the platform, w s the size of stake, & s the time, and i the influence system function. The idea behind quadratc voting is that entites can exercise greater influence but at quadratically ncreasing cost. This changes form one-person-one-vote 1o one person can have more than one equivalent vote but at a higher and higher cost. The fundamental Timiting assumption of quadratic voting, however, is that in general, entiies are not allowed to split their participarion across mulrple pseudonymous identite. Typically the assumpion is that an entit’s participation s steongly fid to ther lga identity. Tndeed thi typically means i it stongly tied to a natural person. Tn oder to use such system the platform must trace and enforce partciparion via strong legal identit

is. As previously discussed, use of an entit's legal identit, may be problemati from privacy perspective for online plaforms if no cost and complexity. When an enity is allowe to spit their stake across muliple participants then quadrati voring has the opposite effect o it intent. Indeed, it strongly incentivizes splitting and has the counter effect of enabling greater influence by an entity for a given available cost or stake. Suppose an entty has some total amount of stake to split among M participants. The tota influence of that entity as the sum of the influence ofthe split panticipants i as follows:

s u i y 2(0)= 2" v =Y =X (o) 2 8)

Consider the following example. Suppose an entity has 50 SoV’ tokens availabl to stake in order to influence some governance feature of the platform. f the entitysplis their stake into two participants with 25 cach their total influence is

(25)2+(25) 2 =5+5=10 a9 Ifinstead the entity does not split, its total influence is, (20) (50)2 =7.071 This demonstrates that a bare or naive inverse quadratic influence function actualy rewards spliting. Indeed the entity may maximize its total inflaence by spliting ts stake into as many participants s is practical. This means that in order for the plarform must enforce a prohibition on spliting via unique legal identicy proving to actualy realize the intent of quadratic VolingThis is 2 problem when we wan to nonlincarly limit the influence of a single entity but foster privacy via consistent preudonymous participant identitis.

One sofuion would be to add a time component to the influence intensity function. The longer the stake is held by a given partcipant the higher the influence. Or even the longer any stake has been held by a given partcipant the higher the influence of all its stake. Thereby an entit is incentivized to add stake built up over fime t its oldest participant identites and not split it with new ones,

Ancther solution is o use nterest o balance influence in a staking based influence intensity system. Tn the case where an entty is engaged on the platform in purchasing products and services, the rewards program can be used to incentivize the concentration of stake into fewer participants. Consequenty there would be 2 cost to an enity to increasing ts tofal influence by splittng its available stake amongst multiple participants. In the case where an entity is not partcipating in the purchase of products or services, there is o balancing inerest to penalize the split

ing of is stake to maximize influence, Some other balancing interest(s) would then be required to penalize spliting. The particulr type of partcipation might provide a way to create balancing interes(s). For example, some platforms use decentralized nereork nodes to provide distributed consensus or validarion or observation of transactions. Node selection could be based on influence intensity a a inverse nonlinear function of a stake such as quadratic voting. I an entity wished to exercise influence over network nodes, that entity would have fo participate in the staking program. If the number of nodes i restricted or has a minimu valu then this forces concentration of stake 10 a limited mumber of selectable nodes or to batter ensure selection. Should those nodes also receive resvards for their performance activity. Then the sze of those rewards could also be a function ofthe actual stake. The higher the stake the higher the reward. This would incentivize an entity who wishes to exercise greater influence over the platform by hosting multiple

Tokenomics White Paper 1 nodes 1o pay a price in lower total rewards and/or risk losing out i the selection process as a result of spliting its stake across ‘multiple nodes. Thus concentration of stzke is required both to be sclected as 4 node and to reccive & higher roward. This incentivizes the concentration of siake despite an inverse quadratic influence intensity function. These staking mechanisms provide baluncing interests o the influcnce.

et another solution is to make the influcnce intensity # function of the activity of @ participant on the platform. This would expliitly link influence 10 interest. 1t would also incentivize node controllers 10 also purchase products and services on the platform. A type of "cat your own dog food". This provides another positive fecdback loop for growth o the platform. As previously described above, when an inverse quadratic function of & quantity such as stake is used (0 directly determine influcnce intensity it actually incentivizes spliting. In contradistinction, when an inverse quadratic or other nonlincar increasing function (where the rate of increase s docrcasing) of some quantity such as stake is used to detcrmine the rate of influcnce intensity when multplied by et some other quantity then it de-incentivizes spliting. Spliting reduces the rale which when ‘multiplicd by some other quantity will always total less across all the splits. So using activity as the other quantity d- incentivizes spliting while stll incentivizing in u nonlinearly increasing way the holding of more stake to gain influence in order o control & node. This is subllc distinction and is why using an inverse quadratic fumction to determine the reward rate de- incentivizes spliting (see section 4.9) but using an inverse quadratic to directly determine a reward or influence does not.

‘These approaches of balancing influcnce and interest also work well with delegated stake or actvity where a participant who docs ot have cnough siake on its own to be selected as a node can reccive delogated stake or delegated activity or both from others. The resultant combination can have higher total influence with higher tolal rewards but both the rewards and o some extent the influcnce arc distributed across the delegatc and delegators. The delcgators have limited control over the node but in combination enable an otherwise weakly staked participant to gamer enough inluence 1o be selected s a node. The controlling entity must be responsive somewhal to t

he influcnce of its delcgators lest they switch their voles 10 a different node. An entity attempting to control multiple nodes must compete with nodes that receive delegated stake. IF it splits its stake across multiple nodes it may not have cnough stake (0 win selection of even a single node. 1f on the other hand the ety splis its stake to join with delegates on muldple nodes it weakens its influcnce over any given node and depending on the roward distribution fumction ‘may or may not reduce the tolal reward it receives,

For cxample, suppose that any sclccted node has a princip controlling entity. This cniity i the delegatc and its node reccives the delegated slake for the purpose of selection. Suppose as well that the total reward increases at a nonlinearly decreasing rate as a function of total sake. Suppose further that the delegate only reccives nonlincar reward at the level of its own stake not as sharc of the total reward duc to the fotal stake of the delcgate and delcgators. Whercas the delegators receive a reward as a ‘nonlingar weighted share of the remaining total reward due (0 the total stake. In this case the delegate risks getting less roward for a given stake even though total reward s higher du to the higher total delegated stake. The delcgate is thereby incentivized o personally stake more und may chose 1o not accept any delegated siake if ts own contribution is high enough to win selection. A delegate s principal cntity has a lot of control over the node but must share the reward with its delcgators. Whercas a delegator has itle control over the node to which it delegates ts stake but receives a higher reward relative (0 ifs delegated siake due 10 the a higher total delcgated stake. If a delcgator concentrates its stake on one node s roward may be higher but it may be less likely to pick a winner whercas distributing its steke across multiple nodes may have the opposite cffect. This balanced approach incentivizes a high degree of concentrated participation and therefore reputable influcnce using only economic rewards without requiring the cnforcement of legal identity proving. OF course a hybrid systemn that uses both might be more appropriate for higher value participation.

1. ConcLusioN By scparating the concerns of medium-of-exchange and store-of-value a dual uilty tokenomics mode has the potential o fix the broken single utilty token model. The store-of-value token is novw relatively unconstrained by the demands of stability, low

iction and high throughput. This enables the store-of-value token (o be designed with beneficial properties that incentivize virtuous participation on the platform that include a side cflect of appreciative valuc. This approach is gencral and can be beneficially adapted to many utility token applications. Samucl M. Smith Ph.D. Is the CTO and co-founder of Difuon, a staraup that is building a decentalized fog computing platform. He is advisory architcct for reputation and AL for ConsenSys. He serves on the technical govemance board of Sovrin.org. Samuel received a Ph.D. in Electrical and Compuer Engineering from Brigham Young University in 1991. He then spent 10 years at Florida Alantic University, eventually reaching full professor satus. His rescarch specialtics include automated reasoning, machine leaming, end autonomous vehicle systems. He has over 100 refereed publications in these areas and was principal investigator on numerous

federally funded research projects. Dr. Smith hus been an sctive participant in open standurds development for networking

protocols and is aserial entrepreneur.

REFERENCES 117 Dusl Token St OF Uity And Security Token-Leverging The e O B Token Word,” Mt Helh 201811315 i e oa-dual-oken seucare Ty and securiy-okens eeraging e bt of bt ok works

Tokenomics White Paper 1s 121°A Glabaly Insroperable Bockchaln or Ity veres e, 31 Robespiere, “Asset Backee Tokens will e 0 the frst ey vausble Uity Tokens.” Wediar, 201805/06 s com @ RobespierreBraniuset backed-tokens il ead-o-he st sy valiable- sy tokers 651 S6bS0erl3 1A Fvams, “On Value, Velocity and Monetary Theors” Blck Chaael, 201801115 s comiblackehannelion vale elocty.andmondir Heory.-new-approsch 0-ryplosset. valutions. 32c9h226 AT 51, Molé, “Token i #2: our Complte Guide f Secuty Token Fxchanges,” Medior, s blog ncufund xgoken-up-2yourcomplte-uids 1o secury-token exchanges-fbAShasdSfe 614 Miano, “Quanstamp Unde Fire Busers Say Faih Shaken n 65 Villon Token,” CobuDesk, 20106/13 st ik comGuantamp-fs-buyer-say-a-shaken-65-milion ok 1714 Lamnguist, “Today's Crypto st Valuation Framevtorks,” BloekChain, 201503 g blockchainaterkely blog odaya-crypio-ssctvaaton-famesroks- 7323 8eda2e 19165, Sinieasan, “Thoughes o Token,” Hedm, 20170527 s cam comhoughte-on-tokens-4361 9 1918 Bemstcin,"Cryptocurencies e money, ot cauiy” Token Econony, 201807722 s keneconamy coeryplocuTeRSics Ars oney-not oy OIS 107 Ty and Sel Digial Currncies,” Sormghd, Hipsisuonghold.n 1111 Lo, TC0s Wers 45% OF IPOS n Q2 2018, As rsptos Disnupt Investment Banks,” Frtes, s e forhos o caiilong 201 K122 cos Wete 450 pos. - 220 Rs-erypios-dsrpt invesament barks HG496 26794 121 Chiny Barers, “Tokens: nvestment Vehicle or Modium of Fxchange (Not Both),” WIT Crypio Economics Lo, 20150713 1315, P, Chaudar, G. G P Prker and M. Van Algsc, Plaform scele: How an emerging business model helps saraps i aege mpies ith i inwestment,” ltfon Thinking Lab, 2015 1141 Bunisk, Crpoasse Vauaions,” fediun, 2017,09:24 eps/medium com ebumiske erpassct valuatons K347 151 Catalind and 1. s, "l Coin Offrngs and th Vel of Crypta Tokens ™ SSRY, vol MIT Sloan Rescarch Paper No. 347.18, 20150313 1161, Comben, “Have You Evr Wonderod Wht Realy Goss nto KYCIAML? 201310726 ips/comcentral comvcyeaml 117 . Masters, "EOS Leads in Tansaction Volunes Abesd of Conptitors.” Cryptovet, vol 201810802, tpsryplovest cominewsicos-blows.al-blockchais-utof the valer-wf-xtaondiary-rasacton-load 118]C. Harper, “In Search of Stabily: An Overviw of the Bling Ssblocon Ecosystm,” itoin Magusie, [ ——— - —— "t campaign-Weckly+Bite-Octohert9%2C+ 20155, SBS ponsored-by s BeMyApDYSD+42APQTL A0 msdium=emlls. ke=ey OFIBWFRBCIAL CIYWI AGHISCZPoaW VLN hSIICIBFO[h2 1Y WSSX2jog g TNDIGRy 9t soursc=Bicoi Magasine-Newslete-2.0 1191, Floyd, “Tezos nvetors G

Chanee o Sl This Weck And They Took I” ColuDes, 201 KT 15 /e coindeskcomiczos ivesors ot chanceo-sell s eck and Ahe-tack i 1201, Parik, “On Tokenomies and 1CO Vaations,” Mediam, 20101113 s/ medium, comideanpatrice 63570/on-okcnomics-and-co-vaharons-S312Shde2bd 121]“Decentalzed écatfices (DIDs)” W3C Drak Community Growp Report 23 August 018, s ceg gt odidspec! 122]“Decentalzed ésatifices (DIDs)”Blocksiackor, 2018109 tpsidoc blockstackongcore ming did bl 1231, Rhodes, “Wh are So Muny 1COs Faiing?,” CoinCentral, 2018104128 Htps/comcenial comy-are.so-rany-seo- g 24] Gamek yos, L AM HODLING," Bitoin Forum, 01312118 s ol ongimden S opic-3756430 25 “Greses Dol ey, Wik, psfen wikipedisrgwikiGreate_ool_tbecry 1261 "Greaer Foo Theory” nestopedia, tpsreinvestopedit.comtermlg grsterolhoorysp 27 “Havvens o st solves theproblemof volatlity i the exyptocconoms:”Cryptainas, 201712718 e cryptonins 201 T/1211 8 hsvvens-dosHoken-sysemsolves-probls-volslty-cryto-cconomy ! 28] HodL” Wikipeds, tptienwkipediaorgwitodl 129)How fo Lentily UBOs: Ultimate ene il Owoers,” oo, 01810424 s o comblogideatyaltmate beneGiabowaces 30]°1C0 Date s Kasse™ 1CO Data, o icodatafoas 2018 3411 duemp o e cumped o work. Lo shovid egultors espond e matre of e bt Fider, Hapsv Gndercom aubusion-colege stuly-co-pump o umps-re- e st option [32)3, Curtier, “The Network Eects Manua: 13 Difernt Network EfTcts und couning)” Medians, 20180109 e com G he networkfTcts-manual- |3 iferent network.fTct il counling-a3e0Th2301 70 [3917, Currier and . . X. Team, “The Networks et Bibl,” nf com, gl corpostaetworkcefectsbible 134]3. Claton,“Sttementon Cryplocurences nd loital Con Offerings.” SEC.gov, s goviaewspublic-slcmen salemen-claylon 20171211 135, Bigs, “Thousands of cyplocrency projets e sleady dend.” fohCrunch, Tt tcherinch coms 201510629 bousands-of ryplocumency-prejects v sieady-dead 36, Pefler, “Doubls ubout the Long-Tecm Visbily o Uiy Crspousses Medium, 201810413 atpsnem, oo p e doublsbout-be-lung - viability- ity -ryploasels-Jb0435051 155 37 T s i buibding an open source decentlized ity ecosystem Do people, ongaizations, apps, and devices.” Decentrlized deniy Foundarion i, Hapdeniy Foundation [T Sedgick, 6% of Lt Year's 1008 Have Faild Alrads.”bicoin com, 20151273 e bicoin com 4.l year cos il syt 3975. P Lalley and . G. Weyl, “Quadratc voting: How mechnisn desipncan adicalive democracy” v, AFA Paprs nd Proceadings 108, pp. 33.37, 2018 40T Neund, 17 you net i s token, o may b Villing e projct you fund (1"

Mediom, 201711107

Tokenomics White Paper 16 s blog ncufund cgi-you-nvestin-spp-okens-sou my-be-Killin th-rectyouund-1-Ghd 1 ShI0F#Se TN, Toman, “On Token Value,”Medium, 2017/08106 Hapsthecontrol.coon-token value <6 1056175 42] . Ml “Blockchuin an te Drcken Revenue Molel: Compleites o Secuiis Lnws and Blockehsin Medium, 2015106111 9] "Open ety System for the Decertrlzed Web," ort e, [ 4] GG, Parker, M. W Vin Alsyne and S. T Chowry, “Pltform Revoltion: How Networked Mrkets Are Transforming the Econurmyand Tov 1o Make Them Work for You,” WW Noron & Comgany, 2016 451 "POA” p vk, ipsipon networkovernance 6] Posnor and . . Weyl, “Radical Mrkets Uprooting Capislism and Democracy for Just Society:” Pineton Usiveesity Pess, 2015 7] “Price Graghs,” ipsfoommarketcapcom 481, . Proki and .. Manolaki, “Iteoducionto Digie Sgnal Procssing. 1988,” Macrillan Publihing Carmpan,pp. 59-62 491 oo afsuthorit” Wikpedis psenwikipedisorgkiTront of authorty 50) “Quadeatic votin.” Wikpedia, Hapsenwikipedia orgwiki Quedetic_ving [511R. R OLesry, “iberem's Grovwing G Crisis (And What's Bein Done 0 Sop 10.” CoinDesk, 201810706 g coindek.comebereus arowing.gas-cisis and i being-done 4o-S0p-1 [52]R, Rottgen, “In Defense of tlty Tokens.”HecerNoon, 20112104 Tt ckermoon comindfense of il lokens- 9711 15975, M. Smith, “Open Reputation Framework.” v, Version .2, 201510513 it comsSenSamue M Papers loh e itepapes apen eputtion Jow el whilepape pif. 15475, M. Smith, “Decentalzed Aviononic Data (DAD) e the hree s of Key Vinagerment,” Spring 2015 it com e OTrustlfosbooting the wh-of st spring 01§ ot Tl documents Decentized AutonscDats pf 15575, M. Smith and D Khavratovie, “Tdenity System Fesenil,” 2013729 1561 “Send and Receve Maney” Stellr, e stellaog developeryguiesigr siradsnsactons il 15715, Comuay, A H. M. M st o *A DID for Everything,” RHOT Pl 2013, 201809726 gt com Ve OTstioirwot? 15515, Cors, “Blockchai Powered Paforms: Tokens” Medium, 2017/10/12 Tapsstoriesplatformesigntoolii comblockchain-powered-plaifors okens- 363755483 599, Foundacon,“Sove: ity For ALL” nipsovrin oy 60]“Sovrin: A Proocol and Token or Sel-Soveren il snd Decentrlized Trust,” Sovrinre, 20180101 tpovrin.urgap-contentiploads 20 5103 Sovtin Proocol.andToken-While Papepdf 617 Sclabelle, “Wha's The Backstory On The W, HODL?,” Medior, 2017104730 s comGiellabelas the-bckatory.cn the word ol 277563925695 16215, Aggarwl, “Undensnding Ether vs Gas.” Mo, 201710627 i com sy Tundersanding cihr v a8 R2620106560 169 “Superposion principle,” Wkipedi, npi

on wiipedisorg/sikiSuperpsin_principle 641 Sy atack,” Wakpedie, ipsenwikipedisorghokiSybil_asek 1651 Koffman, “Vour Officia Guide o he Security Token Esosysen,” Medam, 201 04115 npsmediam com@iarakofTmaniyour-offial guide-o e secuity-token-<casysam- 61850567 1661 Parson, “What i Sccuiy Token? A Comprchensive Gude o Hors They Wk and Thei Impact” HackerNaon, 201511015 Tt iackemoon o whal -5 ecucity-toke--comprehensive- - bow-tre-work s i mpact ¢E9TTad63 [67)“VeCuin Maianet Louncies” CCY, 20180630 Hatpsvcc comivectuin hor blockchain aunches oken-swap-o-ke place il uly 687V Duteri, “On Medium of Exchange Token Valustios, Fait, 20110117 Tl cugeners]201 V107 Timor il 1697 Butriand 6. Weyl, “Liberaion Through Radica Decentslzaion,” M Cryptocurrency, 20181521 Tnpsmeium com GVialButerinherstion-Geh radica decentalizaton 2eAbedc2ac 707 “Why Most 100’ Wll Fal: A Cold Hard Toth.” BockGeeks, npsblockgeeks compideywhy-mostcos wil ol [717W. Hinman, “Digial Aset Transactons: When Howy Ve Gary (Plaic)” SEC.gov, 20106714 sk govinewsspesehispecch-hinman- 061418 721, Mougayar, “Tokenomics A Businoss Guide t Token Usag, Uity and Vel Siartup Manageens 201T06/10 pamapmanagement.org 2017106 10okenormics - busines-guid.fo-token-usags ity v