A MakerDAO Case Study
Disclaimer: Vision Hill is indirectly invested in MKR. The content provided herein should not be considered investment advice, and is not a recommendation of, or an offer to sell or solicitation of an offer to buy, any particular security, strategy, or investment product. Vision Hill has not received and will not be receiving any compensation as a result of this report.
Executive Summary: MakerDAO
MakerDAO is a smart contract platform powered by Ethereum that functions as the first ever purely software-based, community owned and operated decentralized margin facility. In traditional finance, margin refers to the buying of securities with cash borrowed from a broker, using the purchased securities as collateral.
Launched in December 2017, the MakerDAO network offers secured loans of equal cost to anyone in the world through the mechanics of a decentralized autonomous organization (“DAO”). A DAO is an organization represented by rules encoded as a computer program that is transparent, controlled by users and not influenced by a central governing body. The loans in the MakerDAO system must retain at least 2/3rd of the value of the underlying collateral, in other words requiring at a maximum a 66.7% loan-to-value (“LTV”) ratio or a minimum 150% collateralization ratio at all times in order to avoid default. In other words, $1.50 of collateral allows for a $1.00 loan to be issued to borrowers; if the borrower’s collateral drops in value such that these covenants are breached, then some of the borrower’s collateral is automatically sold to garner proceeds that can partially pay down the loan (discussed in greater detail later herein).
The MakerDAO platform backs and stabilizes the value of Dai, MakerDAO’s native collateral-backed stablecoin that is soft-pegged to 1 USD and backed by Ether, the native digital asset of the Ethereum blockchain, as collateral. Dai operates as the by-product of loan generation (or margin), through a dynamic system of collateralized debt positions, autonomous feedback mechanisms, and appropriately incentivized external actors. To get a loan, borrowers first create a collateralized debt position (“CDP”) by posting Ether as collateral through an Ethereum transaction. The collateral of the outstanding Dai is conveniently held in a single pooled Ether contract (“PETH”) and currently includes more than 1.9 million Ether, or just under 2% of the total Ether supply. Much like how government currencies were once backed by precious metals, Dai owners can redeem their tokens for collateral of equal value. It is this transparency that creates trust in the decentralized system; the code and collateral of the MakerDAO system can be inspected and audited anytime, anywhere by anybody that is suspicious about the network and its permissionless margin facility.
Presently Ether is the only collateral accepted in the MakerDAO system, but the platform is expected to accept other assets as collateral soon. Upon the creation of a CDP, borrowers then issue themselves a Dai-denominated loan against the Ether each borrower posted as collateral at a globally standardized rate that currently sits at 0.5% (otherwise known as the “stability fee”), denominated in Dai but paid in MKR (discussed below). Similar to a line of credit, borrowers can withdraw Dai from their CDP at any time, in whole or in part, however, it should be noted that Maker operates more as a margin facility than a credit facility. This debt effectively locks the deposited collateral assets inside the CDP until the outstanding balance is repaid. New Dai tokens are created when borrowers issue themselves a Dai-denominated loan, and these Dai tokens are subsequently burned when repayment occurs. The contracted Ether collateral is returned when the borrowers repay the balance of the Dai that they withdrew, along with the stability fee, in the form of “wrapped Ether” (“WETH”) that can subsequently be converted back to Ether.
According to Placeholder, stability fees (the cost of a loan) have ranged from 0.5% to 2.5% per year, “placing Maker’s credit facility at 1/10th the cost of secured loans offered by traditional financial institutions”. It should be noted that the aforementioned cost comparison is exclusively done on an interest rate basis; other costs such as the direct cost of the Ether collateral (that must already be owned to create a CDP) and the opportunity cost of capital for that Ether are not captured here. With traditional financial institutions, loans can be issued without a borrower having to acquire and post collateral. However, the primary use case of Dai-denominated secured loans thus far has been for users that are sitting on Ether and want to access short-term spending liquidity without having to sell Ether that not only triggers a short-term taxable event but also causes the user to give up the long-term upside potential Ether may offer. Another use case includes leverage.
In digital asset bear markets, CDP owners tend to add additional collateral to their CDP and refrain from withdrawing large sums; in contrast, in digital asset bull markets, CDP owners have the increased ability to spend (or loan) more. The volatility of the digital asset markets drives users to switch between increasing the underlying collateral or increasing the money supply of the Dai stablecoin. The supply of Dai expands as borrowers create loans (Dai is drawn against Ether collateral), and contracts as loans are paid down or liquidated. However, it should be noted that the supply and demand curves here do not intersect; increased demand for Dai does not in turn incentivize increased supply of Dai.
MakerDAO enables anyone to leverage their Ether to generate Dai on the platform. Once generated, Dai can be used in the same manner as any other digital asset: it can be freely sent to others, used as payments for goods and services, or held as long-term savings. Importantly, the generation of Dai also creates the components needed for a robust decentralized margin trading platform. Maker (MKR) is the utility, governance and recapitalization token of MakerDAO. MKR holders set the loan parameters governing the maximum amount of Dai that users can issue themselves. MKR derives its value from the stability fees (e.g., interest payments) and liquidation penalties that are paid by the borrowers. MKR, via its supply issuance algorithm, stabilizes the value of Dai.
In sum, if the value of Dai exceeds $1.00, market participants are incentivized to create more CDPs, resulting in more Dai supply with each DAI worth $1.00 that can be redeemed or traded on the market for greater than $1.00. This increased Dai supply and arbitrage effect stabilizes the price back to $1.00. If, on the other hand, Dai is less than $1.00, it indicates the platform’s value of Ether collateral is trending to be worth less than the amount of Dai it is collectively contracted to back. Keepers (active network participants, described in greater detail later herein) liquidate any “unsafe” CDPs (those that have collateralization ratios < 150%) and auction off the Ether inside these CDPs before the value of the collateralized Ether is less than the amount of Dai it is contracted to back. This Ether is auctioned off for Dai until there is enough Dai to pay back the defaulted debt; this Dai payback is burned (reducing Dai supply) and thus stabilizes the price back to $1.00. Excess Dai is returned to users. Historically the network has proven there has been excess collateral in the CDPs to provide a cushion for the 1–1 USD peg. MKR holders are incentivized to govern the system well and charge a cost of capital that appropriately prices in the risks the network is taking in issuing its loans. This is because if the price movement in Ether or any other form of collateral is volatile enough to cause a shortfall relative to the amount of Dai borrowed against it, the MKR token supply is diluted and sold onto the open market to purchase Dai and offset this imbalance. Simply put: if the Maker network issues bad debt, it is the MKR holders who pay. This ensures the network preserves incentive alignment amongst its users.
Project Background
Much has been written thus far regarding MakerDAO, and we’re certainly not looking to reinvent the wheel with this piece. Rather, the purpose of this piece is to break apart the variables that impact the value of MKR and understand their behavior on both a standalone basis and on a collaborative basis. More comprehensive backgrounds of MakerDAO are available below (in no order of priority), many of which helped this analysis:
Maker for Dummies: A Plain English Explanation of the Dai Stablecoin
Placeholder VC Maker Investment Thesis
Andreessen Horowitz Maker Investment Thesis
Maker Dai: Stable, But Not Scalable
An Elegant Relationship (DAI, ETH, MKR)
Messari Maker Research Profile
Single Collateral Dai — 9 Months in Review
What follows are brief summaries pertaining to Maker’s Project Background, Risk Parameters and Liquidation Procedure from the above sources to set a foundation for our readers before we begin our analysis.
The MakerDAO ecosystem currently has one million outstanding MKR tokens issued, with approximately 72.8% of the total supply distributed and the remaining 27.2% held by the Maker Foundation. The MKR supply will fluctuate based on the performance of the system; new tokens can be created in the rare event that Ether collateral levels decline sharply, and auctions are not sufficient to pay back all outstanding Dai. MakerDAO issued over $200 million in loans since its launch, meaning over 200 million Dai has been drawn against collateral since December 2017. Currently approximately $76 million worth of issued loans are outstanding (current Dai supply of 76 million).
For purposes of our analysis, it is important to highlight a few terms. First up is the Target Price. The Dai Target Price has two primary functions on the Maker platform: 1) it is used to calculate the collateral-to-debt ratio of a CDP, and 2) it is used to determine the value of collateral assets Dai holders receive in the case of a global settlement. The Target Price is initially denominated in USD and starts at 1, translating to a 1:1 USD soft peg. Other terms related to this include Target Rate Feedback Mechanism, Sensitivity Parameter, and Global Settlement, all of which are not covered in this piece but are thoroughly described in the MakerDAO Whitepaper.
Last but not least, it should be noted that the Maker Platform derives its internal prices for collateral and the market price of Dai from a decentralized oracle infrastructure, consisting of a wide set of individual oracle nodes. These oracles provide real time information about the market price of the assets used as collateral in the CDPs in order to know when to trigger liquidations, as well as information about the market price of Dai and its derivation from the Target Rate in order to adjust the Target Rate in certain circumstances. MKR voters choose a set of trusted oracles to feed this information to the Maker Platform through Ethereum transactions, as well as control how many nodes are in the set of trusted oracles, and who those nodes are. According to the MakerDAO Whitepaper, there are certain protections in place that combat against an attacker who gains control of a majority of the oracles, and from other forms of collusion. In fact, according to MakerDAO, up to half of the oracles can be compromised or malfunction without causing a disruption to the continued safe operation of the system.
Risk Parameters
According to the MakerDAO Whitepaper, CDPs have multiple Risk Parameters that enforce how they can be used. Each CDP type has its own unique set of Risk Parameters, and these parameters are determined based on the risk profile of the collateral used by the CDP type. These parameters are directly controlled by MKR holders through voting, with one MKR giving its holder one vote. They key Risk Parameters for CDPs are:
1) Debt Ceiling: The Debt Ceiling is the maximum amount of debt that can be created by a single type of CDP. Once enough debt has been created by a CDP of any given type, it becomes impossible to create more unless existing CDPs are closed. The debt ceiling is used to ensure sufficient diversification of the collateral portfolio.
2) Liquidation Ratio: The Liquidation Ratio is the collateral-to-debt ratio at which a CDP becomes vulnerable to Liquidation. A low Liquidation Ratio means MKR voters expect low price volatility of the collateral, while a high Liquidation Ratio means high volatility is expected. Keepers (described in greater detail later herein) do the work of enforcing this Liquidation Ratio that currently sits at 150%.
3) Liquidation Penalty: Currently 13% This means that if the collateral-to-debt ratio of a CDP falls below 150%, it becomes vulnerable to liquidation. If liquidation is triggered, the outstanding debt of the CDP increases by 13%, and the CDP is then settled based on the price feed. Proceeds from penalty fees are transferred to the PETH pool that increases the ratio of WETH that users receive when they remove their collateral from a CDP. This fee inflates the value of the collateral pool during periods of instability in the market.
4) Stability Fee: The Stability Fee is a fee paid by every CDP. It is an annual percentage yield that is calculated on top of the existing debt of the CDP and has to be paid by the CDP user. The Stability Fee is denominated in Dai, but can only be paid using the MKR token. The amount of MKR that has to be paid is calculated based on a Price Feed of the MKR market price. When paid, the MKR is burned, permanently removing it from the supply.
5) Penalty Ratio: The Penalty Ratio is used to determined the maximum amount of Dai raised from a Liquidation Auction that is used to buy up and remove MKR from the supply, with excess collateral getting returned to the CDP user who owned the CDP prior to its liquidation. The Penalty Ratio is used to cover the inefficiency of the liquidation mechanism. During the phase of Single-Collateral Dai, the liquidation penalty goes to buy and burn PETH, benefiting the PETH to ETH ratio.
As noted above, MKR holders vote on Active Proposals (Single Action Proposal Contracts and Delegating Proposal Contracts, both of which are discussed in greater detail in the whitepaper) to modify the internal governance variables of the Maker Platform.
Liquidation Procedure
To ensure there is sufficient collateral in the system to guarantee the value of all outstanding Dai, a CDP can be liquidated if it is deemed to be too risky. This occurs when the value of the collateral (as judged by the oracles) is less than the required collateralization for the CDP. As noted previously, if a borrower’s Ether collateral drops in value such that the 2/3rd LTV covenant (or 150% collateralization ratio) is breached, then enough collateral is seized and automatically sold to repurchase Dai and partially pay down the loan, covering the outstanding debt plus fees. The remaining collateral is then made available to the user (the CDP owner). Active network participants with self-economic interests called “Keepers” operate as external arbitragers and do the work of enforcing this liquidation process, and have made hundreds of thousands of dollars in nearly risk-free profit in the last several months. A Keeper is an independent (usually automated) actor that is incentivized by profit opportunities to contribute to decentralized systems. Profit opportunities are presented whereby Keepers can buy PETH from a liquidating CDP at a 3% discount to the spot (or market) price, and turn around and sell the PETH on the market at the spot or market price, all in a single atomic transaction. The execution of these arbitrage trades typically takes a few minutes, thus Keepers do incur some market risk in the process, albeit relatively minimal. In sum, Keepers help maintain the health of the entire MakerDAO ecosystem by ensuring that unsafe debt positions are offered up for liquidation as quickly as possible. This is particularly important during highly volatile market drawdowns as collateral value could fail to satisfy the debt obligation.
The order of operations during liquidation for single-collateral Dai is as follows (discussed in greater detail in the MakerDAO Whitepaper):
1) The defaulted CDP is closed by a Keeper that sends the CDPs assets to a Liquidity Providing Contract (“LPC”), which in turn, offers the CDP’s assets for sale. The LPC is simply a smart contract that trades directly with Ethereum users and Keepers according to the price feed of the system.
2) The 13% liquidation penalty is applied to the Dai-denominated loan in addition to the stability fee.
3) The LPC removes PETH collateral to satisfy the outstanding debt at current oracle prices.
4) The CDP owner is now able to remove their remaining collateral from the closed position (it is our outstanding this is done in the form of WETH that can subsequently be converted back to Ether). The CDP owner receives the value of the leftover collateral minus the debt, stability fee and the liquidation penalty.
5) The seized PETH is set for sale in the LPC at dai.markerdao.com with an incentivizing discount for Keepers (currently 3%) applied to the value. As noted previously, Keepers can atomically purchase the PETH by paying Dai.
6) The Dai earned from the sale of PETH is burned to wipe out the CDP debt. In other words, all Dai paid this way are immediately removed from the Dai supply, until an amount equal to the CDP debt has been removed.
7) If there is excess Dai from the sale, it is sold for PETH that is in turn returned to the Pool of Ether, inflating its value. In other words, if any Dai is paid in excess of the debt shortfall, the excess Dai is used to purchase PETH from the market and burn it. This positively changes the ETH-to-PETH ratio and results in a net value gain for PETH holders.
8) If there is insufficient Dai from the sale, then PETH is drawn from the pool and offered for sale to cover the shortfall. This dilutes the total value of the pool. In other words, if the PETH selloff initially does not raise enough Dai to cover the entire debt shortfall, more PETH is continuously created and sold off. New PETH created this way negatively changes the ETH-to-PETH ratio and results in a net value loss for PETH holders.
In sum, for Single-Collateral Dai, the Maker Platform determines when to liquidate a CDP by comparing the Liquidation Ratio with the current collateral-to-debt ratio of the CDP. Each CDP type has its own unique Liquidation Ratio that is controlled by MKR voters and established based on the risk profile of the particular collateral asset of that CDP type. Liquidation thus occurs when a CDP hits its Liquidation Ratio. The Maker Platform will automatically buy the collateral of the CDP and subsequently sell it off.
An auction model will replace the current mechanism when the system transitions to Multi-Collateral Dai. According to the MakerDAO Whitepaper, during a liquidation, the Maker platform will buy the collateral of a CDP and subsequently sell it in an automatic auction. This auction mechanism will enable the system to settle CDPs even when price information is unavailable. In order to take over the collateral of the CDP so that it can be sold, the system first needs to raise enough Dai to cover the CDP’s debt. This is called a Debt Auction, and will work by diluting the supply of the MKR token and selling it to bidders in an auction format. In parallel, the collateral of the CDP will be sold in a Collateral Auction where all proceeds (also denominated in Dai) up to the CDP debt amount plus a liquidation penalty (A Risk Parameter determined by MKR voting) will be used to buy MKR and remove it from the supply. This will directly counteract the MKR dilution that happened during the Debt Auction. If enough Dai is bid to fully cover the CDP debt plus the liquidation penalty, the Collateral Auction will switch to a reverse auction mechanism and try to sell as little collateral as possible. Any leftover collateral will be returned to the original owner of the CDP. Keepers will participate in the Debt Auctions and Collateral Auctions when CDPs are liquidated.
Modeling MakerDAO
As touched on previously, holders of MKR tokens play several important roles in the MakerDAO system from voting on proposals to modify the internal governance variables of the network, to choosing a set of trusted oracles as well as controlling how many nodes are in the net of trusted oracles and who those nodes are; and by directly controlling (via voting) the Risk Parameters (described previously), including the Liquidation Ratio. The incentives of MKR holders are well-aligned with the interests of Dai holders, because the economics of the MKR token are such that those who hold it are financially rewarded for commensurate with the growth of Dai. The price of the MKR token, like most traditional assets, is dictated by the forces of supply and demand.
Before we begin, we want to note to readers that what follows is not a valuation approach. We are not performing a discounted cash flow analysis or otherwise trying to value MKR with traditional methodologies. Rather, what we are attempting to do, is examine the value-driving variables of the MakerDAO network all in isolation of each other and explore the price impact on MKR if we make an incremental unit change to each variable. What inspired our analysis is the excellent work product by Aviv Milner and Victor Hogrefe in their Investment Analysis: MakerDAO Medium post from May 2018. Our analysis below builds off of their fine work and discoveries, and we are not attempting to reinvent the wheel here. For a deeper understanding of the mathematical applications and intra-variable relationships as well as their corresponding impacts on MKR’s price, we suggest reviewing Milner and Hogrefe’s work in more detail.
This analysis is segmented into three sections. In Section I, we conduct a variety of intra-variable sensitivities and examine their price impact on MKR in isolation. These variables include the burn rate per annum, the stability fee (otherwise referred to as the governance fee), and time. Let’s define each variable upfront:
i) Burn Rate — This is the annual burn rate of MKR as a percentage of the network. A noteworthy takeaway from Milner and Hogrefe’s findings is that the stability fees burned are not dependent on the price of MKR; if the price of MKR is very high, users will burn smaller amounts and vice versa. Thus, the stability fee burned is irrespective of the MKR price. Additionally, as Marc-André Dumas shares in these series of tweets, the liquidation penalties generate a substantial portion of the revenues generated by MakerDAO, in fact much more than the stability fees. Dumas notes this penalty percentage is expected to trend down over time as Multi-Collateral Dai comes to market, but nonetheless we believe this is an important point to acknowledge.
ii) Stability Fee — As noted previously, MakerDAO sets a settlement stability fee in the form of an annual percentage (currently 0.5% of outstanding Dai per year) that is required for users to retrieve the collateral locked away. This can inherently be viewed as a governance fee as it is collected (via burning) by MKR holders, incentivizing them to continue modifying the internal governance variables of the MakerDAO network.
iii) Time — In traditional finance, the time value of money is the concept that money available at the present time is worth more than the identical sum in the future due to its potential earning capacity. In MakerDAO’s case, time enables the MKR supply to expand or contract as a result of a varying burn rate, enabling us to separately measure the price impact across different time factors.
The sensitivities that follow are segmented into three parts. Part I analyzes only the burn rate and stability fee in isolation. Part II analyzes what happens if we add back time and measures the return profile from a price perspective. Part III is an expansion of the findings of Part II, but from a supply perspective.
Section II of this report extrapolates the price of MKR under a few conditions as an experiment to explore what MakerDAO could look like if Dai were to undergo aggressive monetization.
Let’s begin.
Section I: Part I
Recall MKR tokens are burned as part of the fee for settling a CDP. Burning acts as a supply contraction mechanism; the burn rate represents the annual rate (as a percentage of the network) of MKR that is burned. In order to sensitize the burn rate, we first had to compute the current estimated burn rate of the MakerDAO network. Using the following inputs, and the mathematical findings of Milner and Hogrefe’s work, we solve for the estimated current burn rate (highlighted in light green) of the network via iteration:
As can be seen above (highlighted in light green), we compute an estimated current burn rate of 0.14%. What we chose to sensitize here is how 1) changing the burn rate in isolation impacts the MKR price, 2) how changing the stability fee in isolation impacts the MKR price, and 3) how changing both these variables together impacts the MKR price. This enables us to strip away (e.g., disconnect) both the time value and, when sensitizing the burn rate only, the governance value associated with MKR. We then explore the value creation/capture with the burning mechanism by itself, and with the stability fee by itself.
Before we continue, we felt it important to analyze the historical estimated burn rate for the MakerDAO network to determine the reasonableness of the current estimated burn rate of 0.14%. As the chart below shows us, historically the network’s burn rate has trended upward if we observe it in a linear lens, with a high of 0.66%, a low of 0.005%, and an average and median of 0.16% and 0.072%, respectively. As can also be seen, the burn rate is heavily influenced by the stability fee that varies from time to time based on community voting. Based on these historical findings, and the current stability fee of 0.5%, our current burn rate estimate of 0.14% appears reasonable.
Being mindful of the current estimated burn rate of 0.14%, and based on the historical burn rate behavior, we created a burn rate range of 0.05% to 0.65% in a series of +0.05% increments. Additionally, given the current stability fee of 0.5%, we created a stability fee range of 0.0% to 1.5% in a series of +0.25% increments. We used a MKR price of $375.53 for this analysis.
What we found is, the higher the stability fee, the higher the MKR price (*all else equal*), but the smaller the increase in MKR price in relative terms (diminishing returns to scale). Each +0.25% increase in stability fee at any constant burn rate initially contributes to a 2x (+100.0%) price increase (if we start stability fees at 0.0%) but then trends down to a 16.7% price increase contribution (at a stability fee of 1.5%). It is important to note that as the price rises, the lower the percentage of MKR’s absolute value is burned when CDPs are closed out.
The blue highlight in Table #1 illustrates the current MKR price of $375.53 we previously solved for with an estimated 0.14% burn rate and a 0.5% stability fee. We next analyze the price impact of an additional +0.25% increase in stability fee at each specified burn rate.
The blue highlights in Table #2 are reference points to help us interpret this table. The first $521 blue output informs us that, with an assumed burn rate of 0.05%, each +0.25% increase in stability fee contributes to a +$521 increase in MKR’s price. This can be found by simply taking the difference between each incremental price point at a 0.05% burn rate in Table #1 (e.g., using a stability fee of 0.75%, and a burn rate of 0.05%, we have a MKR price of $1,564, which is $521 higher than the MKR price of $1,043, computed with a stability fee of 0.50% and a burn rate of 0.05%). The second $261 blue output in Table #2 informs us that, with an assumed burn rate of 0.10%, each +0.25% increase in stability fee contributes to a +$261 increase in MKR’s price. This can be found by simply taking the difference between each incremental price point at a 0.10% burn rate in Table #1 (e.g., using a stability fee of 0.75%, and a burn rate of 0.10%, we have a MKR price of $782, which is $261 higher than the MKR price of $521, computed with a stability fee of 0.50% and a burn rate of 0.10%).
So what can we conclude from this? We conclude that by keeping the burn rate constant, each +0.25% incremental increase in stability fee contributes to the following percentage price increases MKR:
We next analyze the price impact of an additional +0.05% increase in burn rate at each specified stability fee.
The blue highlights in Table #3 are reference points to help us interpret this table. The first ($261) blue output informs us that, with an assumed stability fee of 0.25%, the first +0.05% increase in burn rate contributes to a ($261) decrease in MKR’s price (this might seem paradoxical, but remember, we are ignoring time in this specific exercise). This can be found by simply taking the difference between the first incremental price point at a 0.25% stability fee in Table #1 (e.g., using a stability fee of 0.25%, and a burn rate of 0.10%, we have a MKR price of $261, which is $261 lower than the MKR price of $521, computed with a stability fee of 0.25% and a burn rate of 0.05%).
The second $87 blue output in Table #3 informs us that, with an assumed stability fee of 0.25%, the second +0.05% increase in burn rate contributes to a ($87) decrease in MKR’s price. This can be found by simply taking the difference between the third incremental price point at a 0.25% stability fee in Table #1 (e.g., using a stability fee of 0.25%, and a burn rate of 0.15%, we have a MKR price of $174, which is $87 lower than the MKR price of $261, computed with a stability fee of 0.25% and a burn rate of 0.10%).
So what can we conclude from this? We conclude that by keeping the stability fee constant, each +0.05% incremental increase in burn rate contributes to the following percentage price decreases MKR:
The result of this Part I should seem paradoxical, particularly with respect to the results of our sensitivity in Table #3. If the burn rate increases, and the MKR supply contracts, and thus becomes scarcer as a percentage on absolute dollars, all else equal, the price should expectedly appreciate in value, not decrease in value. Given our sensitivities in this Part I carved out the burn rate and the stability fee in isolation to assess their economic impacts on MKR price, the time factor was ignored (and thus, the MKR supply change as a result of a varying burn rate has not gotten captured into this price impact). In this lens, burn rate is essentially (*all else equal*) the required rate of return the MakerDAO network must produce to justify its current value and otherwise satisfy its economic demand. Milner and Hogrefe’s work demonstrates that the burn rate in this regard can be thought of as a dependent variable on MKR similar to how PE or EV/Sales are dependent variables on stock prices. Thus, the burn rate can be considered a decent (negatively correlated) comparable driven by network utilization and forward-looking confidence. If a MakerDAO competitor existed but was less attractive, the required return to investors would be higher thus require a higher burn rate. We explore adding back time in Part II that follows.
Section I: Part II
In this Part II, we explore the impact on MKR price when we add back time. This enables us to capture the price impact of the MKR supply change as a result of a varying burn rates. We assessed the time value in two intervals here: 1-year and 5-years.
The net result is a return-on-investment (“ROI”) for MKR holders ranging from 0.2% (for lowest burn rate of 0.05%) to 2.6% (for highest burn rate of 0.65%). This makes intuitive sense because, as noted previously, if the burn rate increases, and the MKR supply contracts, and thus becomes scarcer as a percentage on absolute dollars, all else equal, the price should expectedly appreciate in value, which this sensitivity demonstrates.
It should be noted that this sensitivity in Part II explicitly focuses on the price impact of MKR, but doesn’t quite closely examine the supply mechanics. We explore this in Part III later.
In order to add back time, we must compute the MKR supply as adjusted by the burn rate over our established time horizons. Using the mathematical findings of Milner and Hogrefe’s work, we have:
What follows are sensitivities, much like what we did in Part I, but this time we are including time factors that enable the MKR supply to change as a result of a varying burn rate.
What we found here, is that each +0.05% increase in the burn rate translates to the following ROIs on MKR over 1-year and 5-year periods:
As we can see, all else equal, the higher the burn rate, the higher the ROI to MKR holders over both a 1-year and 5-year time horizon. This assumes both the starting MKR network value and the stability fee are held constant throughout the associated time periods, and strictly measures the economic return driven by the burn rate to existing MKR holders. The ROIs in 1-year are calculated by dividing the MKR price in 1-year by the MKR price (calculated with our prior formula) at time zero under the same stability fee and burn rate assumptions. Similarly, the ROIs in 5-years are calculated by dividing the MKR price in 5-years by the MKR price (calculated with our prior formula) at time zero under the same stability fee and burn rate assumptions.
To put these ROIs into context, U.S. 1-year and 5-year nominal Treasuries are currently offered at 2.56% and 2.51%, respectively, suggesting the ROIs to MKR holders under these specific assumptions may look rather low in relative terms. However, it is important to recall that the MakerDAO platform is an overcollateralized margin platform, requiring at a minimum a 150% collateralization ratio as described previously.
Nonetheless, what excites us about MakerDAO is the fact the platform serves as a mechanism that allows for any synthetic asset to be created and pegged to a prescribed value (while being backed by parties with skin in the game). A classic future use case for this could be the refinancing of late-stage mortgages for 0.50%, the cost of the stability fee. In such a circumstance, one could collateralize a tokenized deed and use Dai to pay down the rest of the mortgage. Additionally, Dai has entered beta testing on Wanchain as the first cross-chain ERC20 token as of October 2018. Another particularly interesting development that could arise in the future is if (or rather, when) a credit default swap market develops for MakerDAO CDPs, incentivizing market participants to trade around the credit risk inherent in CDPs.
We next analyze the proportional value creation over these time respective periods.
The value creation in 1-year (column “A”) matches the ROI previously calculated under a single year time horizon. The value creation in years 2–5 (column “B”) is calculated by taking the proportional increase in value of each numerical output in Table #5 and dividing it by the corresponding numerical output in Table #4. Finally, the value creation after 5-years (column “C”) is calculated by summing up columns “A” and “B” together. We can note this column C matches the 5-years ROIs we previously computed.
As noted earlier herein, this Part II analyzed what happens to MKR if we add back time, by measuring the return profile strictly from a price perspective. Part III that follows is an expansion of the findings of this Part II, but examines the mechanics more closely from a supply perspective.
Section I: Part III
In this Part III, what we sought to explore here is, as the MKR supply continues to contract at an equal burn rate over 1-year and over 5-years, what the net economic impact is to MKR holders. What we’ve found in the sensitivities that follow, is that the supply burned, as a percentage of the total supply, gradually increases over time.
In 1-year, a +0.05% burn rate increment removes approximately 363 MKR tokens (depending on the burn rate, it ranges from 362–364) from the current starting supply of 728,228. As a percentage of the current supply of 728,228, the 363 MKR tokens burned equals 0.05%. But what does this look like in 5-years? Is it as simple as just multiplying the number of MKR tokens burned in 1-year (approximately 363) by a factor of 5 and examining that as a percentage of the original total supply?
Not quite.
What the 5-year sensitivity here shows us is the effect of contraction compounding. In relative terms, the percentage of supply burned in 1-year with each +0.05% change in burn rate equates to 0.05% when measured against the prior burn rate’s supply after 1-year. In 5-years, in relative terms, we see the percentage of supply burned ticks up to 0.25% (the original 0.05%, multiplied by a factor of 5), *however*, the supply burned in absolute terms does not equate to the supply burned in 1-year multiplied by a factor of 5. We see this by acknowledging if 363 * 5= 1,815, only two outputs of our 5-year sensitivity analysis display numbers in that range (the rest are significantly lower). This is due to contraction compounding; because the original base (the starting total supply in this example) is becoming smaller and smaller each year due to supply contraction, and as a result, the constant burn % in each year results in a smaller and smaller absolute change to the original base.
So now that we’ve concluded the supply contraction in absolute terms is being reduced as time goes on, we next observe how this impacts MKR holders. If the burn rate increases, and the MKR supply contracts, and thus becomes scarcer as a percentage on absolute dollars, all else equal, the price should expectedly appreciate in value. We measured that price appreciation over 1-year and 5-year time periods and observed the following return profile:
As we can see above, we have a matching return profile compared to our value creation analysis between the two time periods (years 2–5) in Part II.
Now, what happens when we experiment with Dai’s debt ceiling? This leads us to Section II.
Section II
Now that we understand the intra-variables of the MakerDAO network and their isolated economic impacts, we can extrapolate the price of MKR (rather optimistically, for experiment) under a few conditions to visualize what the network could look like if Dai were to undergo aggressive monetization.
In Section I, Part I of this analysis, we computed an estimated current burn rate (per annum) for Maker of approximately 0.14%. As we noted previously, the current Dai ceiling is approximately 76 million. Keeping the burn rate per annum and the stability fee both constant at 0.14% and 0.50% respectively, and assuming total supply, we extrapolate the MKR price forward 8-years (beginning of 2019 through the end of 2026) under two scenarios: 1) the Dai ceiling triples in 0.5 years, and 2) the Dai ceiling triples in 2.5-years. What follows is a chart illustrating the MKR price discovery under these assumptions:
The results this extrapolation exercise suggest, all else equal and under the specific aforementioned assumptions, that MKR can achieve a $1,000 price point between 10-months (upside case) and 4.6-years (downside case). The upside case enables MKR to reach a $5,000 price point in 1.9-years and a $10,000 price point in 2.4-years, while the downside case enables MKR to reach those same price points well after 8-years.
So, what does this tell us? What does this mean for Dai and Ethereum?
In the cases where MKR is priced at $1,000, the Dai ceiling is projected to be approximately 275 million, equating to 0.0004% of global government debt and 0.0013% of U.S. government debt as of the end of 2018. Given MakerDAO’s near-term plans to implement Multi-Collateral Dai, if we over-exuberantly assume Ethereum transitions into a widely accepted digital global reserve asset during this time frame and 100.0% of the Ethereum supply is collateralized (in context, currently about ~1.9% of the Ethereum circulating supply is collateralized), we find that translates into 38.1% of the Dai ceiling being collateralized by Ether (104.7 million ETH supply / 275 million Dai ceiling). Assuming Ether is collateralized at a 150% collateralization ratio, this implies a network value for Ethereum of $16.6 billion (based on current prices) based on MakerDAO’s use case alone, translating to a 1.5x increase over Ethereum’s current network value ($11.1 billion).
If we repeat the same exercise for the $5,000 and $10,000 MKR price points, we derive the same network value for Ethereum of $16.6 billion, but find that (all else equal) that this translates into ~7.3% (for Dai ceiling of 1.4 billion at $5,000 MKR price) and ~3.9% (for Dai ceiling of 2.6 billion at $10,000 MKR price) of the Dai ceilings being collateralized by Ether.
We can recall that a noteworthy takeaway from Milner and Hogrefe’s findings is that the stability fees burned are not dependent on the price of MKR; as the MKR price rises, the lower the percentage of MKR’s absolute value is burned when CDPs are closed out.
In this next extrapolation, we double the time horizon from 0.5–2.5 years to 1–5 years. Taking the Dai ceiling growth estimates over these new time horizons, we experiment (rather optimistically) with varying annual burn rates: 0.14% per annum (the current estimated burn rate), 0.25% per annum and 1.0% per annum. One additional key change we elected to make here is we will use an initial 1 billion Dai as a starting ceiling, rather than the current ceiling of approximately 76 million. The rationale for this is if we view Dai as a viable competitor in the stable coin market, the Dai ceiling will be expected to reach the low single digit billions (Tether currently has over 2 billion in circulating supply). This is simply just done for experimental purposes under the assumption Dai undergoes aggressive monetization and MakerDAO hits critical mass. What follows is a chart illustrating the MKR price discovery under these assumptions:
The results of this multi-scenario extrapolation exercise suggest, all else equal and under the specific aforementioned assumptions, that in the highest Dai ceiling growth environment, MKR can achieve a $10,000 price point at approximately 1.5-years (highest per annum burn rate) and approximately 5.25-years (lowest per annum burn rate). The Dai ceilings in these scenarios range from approximately 2.8 billion (highest per annum burn rate) to approximately 20.2 billion (lowest per annum burn rate). If we once again over-exuberantly assume Ethereum transitions into a widely accepted digital global reserve asset during this time frame and 100.0% of the Ethereum supply is collateralized at a 150% collateralization ratio (for context, currently about ~1.9% of the Ethereum circulating supply is collateralized), we find that (all else equal) this translates into ~3.7% (for Dai ceiling of 2.8 billion at a $10,000 MKR price with the highest burn rate) and ~0.5% (for Dai ceiling of 20.2 billion at a $10,000 MKR price with the lowest burn rate) of the Dai ceilings being collateralized by Ether. As the total supply of Ethereum did not change in this scenario, we derive the same network value for Ethereum of $16.6 billion, which implies a 1.5x increase over Ethereum’s current network value ($11.1 billion).
Repeating this exercise for the lowest ceiling growth environment, only the highest burn rate scenario results in the MKR price reaching $10,000 within the 8-year time horizon (achieved within ~7.25 years). At this level, the Dai ceiling is approximately 2.8 billion, which, under the aforementioned repeat assumptions, and all else equal, translates into ~3.7% of the Dai ceiling being collateralized by Ether.
Grateful thanks to Tom Shaughnessy, Qiao Wang, and especially Tanner Hoban for all their help and valuable feedback with this analysis.
— — — — — — — — —
Vision Hill Advisors is a digital asset focused fund of funds. Through a proprietary fund manager selection process and institutional level due diligence, Vision Hill aims to lead investors into the future of digital assets. Vision Hill brings together a team with extensive experience in traditional financial markets, a deep passion and understanding of digital asset markets, and a history of risk and portfolio management.
Vision Hill is indirectly invested in MKR. The content provided herein should not be considered investment advice, and is not a recommendation of, or an offer to sell or solicitation of an offer to buy, any particular security, strategy, or investment product. Vision Hill has not received and will not be receiving any compensation as a result of this report.
