Chit Fund DAO

The Chit Fund DAO (cfDAO) is a DAO devoted to decentralized banking. The cfDAO generalizes the function of traditional chit funds to make an open global deFi platform for all types of banking. The cfDAO uses DGF to manage its underwriting and governance.

Background

A chit fund is a folk banking scheme traditionally used by people without access to formal banking institutions. The term chit fund comes from India, where it is used by a significant percentage of the population, and has become legally regulated in many regions. Similar schemes are used throughout the world, especially in rural areas without formal banks.

Banking involves two money processes: deposits and withdrawals. The sequence determines the type of banking. Consider for instance, the fundamental banking practices of loans, investments, and insurance. A loan is an initial withdrawal of money followed by a series of deposits. An investment is a series of deposits, followed by a withdrawal. Insurance is a series of deposits, with a withdrawal at random times between the deposits. These three types of banking are unified by the scheme in a basic chit fund, as illustrated next.

Basic chit fund scheme

In a traditional chit fund there are several subscribers and an organizer. As a basic example, suppose 100 subscribers agree to deposit $200 in the fund every month for 100 months. Each month an auction takes place, where the subscribers can bid to receive a fraction of the current total chit fund. The lowest bid wins. Once you win an auction, you can’t win another, but you are still required to add your $200 each month until the end. So the number of subscribers is always equal to the number of months the chit fund runs in traditional chit funds (though this stipulation is removed in the generalized cfDAO).

In this example, notice that $20,000 enters each month for 100 months. For $200 entry you can withdraw almost $20,000 in the first few months, if you bid lowest. Early winning bids will typically be from people eager to use the chit fund as a loan, so they will typically be lower than $20,000. Since the excess money remains in the fund for later bids, the fund typically grows, in which case it holds more than $20,000 later. Therefore, early withdrawal is like taking a loan that is repaid for 100 months with interest. Later withdrawal is like a bank deposit that earns interest. Others can use the fund as insurance by withdrawing when they need it at any time between the beginning and the end.

Web3 improvements

Web3 technology can greatly improve the efficiency and security of the traditional chit fund scheme.

First, smart contracts replace the organizers, eliminating the risk of an organizer absconding with the fund, and decreasing their typical 5% commission.

Second, by implementing a reputational system, honest participation can be tracked across many different chit funds, as people will earn reputation each time they pay their required monthly fee. The more often they participate honestly, the larger their reputation will grow. Then participation in larger and more complex chit funds can be dependent on reputation, which will incentivize healthy collaboration.

It will come as no surprise that one of the major overheads making this scheme less efficient is subscribers defaulting on paying the premia. It’s been estimated that 35% of chit fund subscribers have defaulted at least once recently and 24% have defaulted after winning an auction.  

All things being equal, larger groups dilute the risk of individual default. A chit fund with a large number of people with high reputation can pay small premia for insurance. 1 One million people investing $1 daily for 50 years allows 50 people per day to immediately begin taking an average payout of roughly $20,000. More investors investing smaller amounts in shorter increments, means more people can withdraw at any point. By monitoring and analyzing the performance of a fund, programs can suggest values that can be withdrawn at any given time. By automatically bidding when the fund rises above expected levels, the fund can be stabilized to give predictable returns—especially if multiple funds are connected.

With no initial reserve backing, this allows people to bootstrap their way to greater financial security and stability, assuming the reputation system is sufficiently strong to guarantee a low percentage of defaults. Further efficiency-improving mechanisms are mentioned below, where we discuss a more general architecture for underwriting.

Like other overhead costs (the appeals process, policing, etc.), insurance is cheaper when the system is running well. The more automated the decentralized economy becomes, for example, with code-is-law smart contracts, the less insurance is required. The purpose of insurance is to decentralize risk. Ideally the risk would be decentralized perfectly and then the need for insuring individual transactions would disappear. Given that no system will ever be perfect, insurance will always be essential. But by implementing the new tools of information technology and the new architectures of P2P distributed computing, we can create more effectively decentralized organizations. Decentralizing risk makes insurance more stable and efficient, which improves the economy.

However, the core difficulty in implementing any chit fund is organizing it. The primary threat is that a subscriber will win an early bid, then fail to make the rest of the deposits--defaulting. Traditional chit fund managers make a 5% commission, for the primary goal of ensuring subscribers do not default. This is especially challenging in a decentralized, open, global financial scheme. How do we guarantee the default rate is low enough to justify participation from pseudonymous participants? To solve this problem, the chit fund needs to be insured against the risk. Therefore, the subscribers of the cfDAO require underwriting.

Basic cfDAO function

The cfDAO relies on an underwriting DAO (uDAO).

In the above figure, we illustrate five basic interoperating smart contracts which govern the cfDAO.

In the first module, the chit fund smart contract (cfSC) governs instantiations of the chit fund. Each subscriber $S_{i}$ makes monthly deposits, called premia, into the chit fund. This gives subscribers the right to bid on the collection--called the premia pool. The cfSC keeps the accounting for each round until the chit fund's conclusion.

The second module is the Reserve Smart Contract (rSC), which initiates the underwriting reserve. In this smart contract, underwriters pool money to be held in reserve. This reserve money is used to guarantee chit funds continue to be funded properly in case any subscriber defaults. Each underwriter who deposits cash in the rSC is issued proportional underwriting REP tokens (uREP).

In the third module, underwriters may encumber uREP tokens in order to underwrite chit fund subscribers accepted by the underwriter. In case a chit fund subscriber defaults, then this Underwriting Smart Contract (uSC) will use the uREP tokens to continue the deposits until the chit fund ends. For this service a chit fund subscriber must pay an underwriter a fee. For example, a standard manager's fee in a traditional chit fund is 5%.

In the fourth module, the underwriter DAO follows basic DGF workflow. Instead of keeping the underwriting fee, the underwriter is given uREP, and the fee is distributed to all uDAO members in proportion to their uREP holdings through the REP salary. The uDAO allows the members to police each instantiation of the chit fund through the Validation Pool, to guarantee all subscribers are fully underwritten to ensure successful conclusions. Chit funds which are not fully underwritten are automatically rejected.

Initially the rSC requires underwriters to make 100% capital reserves. However, as detailed in the page on decentralized underwriting, this 100% reserve requirement can be safely diminished once stable fees are proven, because the uREP tokens will gain intrinsic value from the REP salary. A fifth module makes this explicit, with a uREP Marketplace Smart Contract. When a subscriber defaults, the uREP tokens underwriting them are sold at auction until the premia are paid. If the encumbered uREP is sufficient, then the remaining tokens are returned to the underwriter. If the encumbered uREP is insufficient, then new uREP is minted and sold at market to cover the premia. In this case the entire uDAO suffers because their uREP holdings are diluted, which motivates proper policing in the fourth module Validation Pool.

Advantages

Unmanaged interest rates

In this decentralized banking scheme, interest is automatically handled. When more people want to save their money (invest), then interest decreases naturally. When more people want to take loans, interest increases. Therefore, there is no need for central planning for banking rates.

Automated optimal actions

A sophisticated cfDAO will make the numerical decisions automatic. This is because there are optimal actions based on the subscriber's goals. For example, if the subscriber wishes to use the chit fund as an investment, then the proper bid in any round is given by the formula $\${\Bigl (}Np+{\tfrac {F-Np}{T-i+1}}{\Bigr )}$. The notation is explained next with a justification for the formula. Denote the number of rounds in the chit fund as $T$. The current stage is $i$. The premium is $\$p$. The number of subscribers is $N$. (In the basic chit fund $N=T$.) At each stage $i$ the subscribers add $\$Np$ to the fund $\$F_{i}$ so at each stage $\$F_{i}\geq \$Np$. Note that if two or more subscribers have not bid on the fund by the final round, then they split the fund equally.

Assuming the interest rate is positive, $\$x$ is worth more than $\$x$ in the future. Consider the case when there are two bidders who wish to maximize their earnings from the chit fund who are the only ones left to bid in the second to last round. As a rational bidder you should accept $\${\Bigl (}Np+{\tfrac {F-Np}{2}}{\Bigr )}$ on the second to last round, because otherwise you split the same amount later in the final round. So you might as well accept that number earlier. If your opponent underbids you, then you will receive more on the final round. However if you do not bid that low, then your opponent can bid slightly higher than that, diminishing your rewards on the final round. Similar logic applies to the preceding rounds, giving the optimal strategy for an investor is to bid $\${\Bigl (}Np+{\tfrac {F-Np}{T-i+1}}{\Bigr )}$ at stage $i$.

For example, if everyone is going to save money the entire time, it's better to get out of the chit fund if you are planning on investing, in order to join another fund that has members who are seeking loan opportunities. So you should always bid $\$Np$ on the first round and get out if no one bids low.