From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from whitealder.osuosl.org (smtp1.osuosl.org [140.211.166.138]) by lists.linuxfoundation.org (Postfix) with ESMTP id A858CC016F for ; Tue, 21 Jul 2020 03:40:23 +0000 (UTC) Received: from localhost (localhost [127.0.0.1]) by whitealder.osuosl.org (Postfix) with ESMTP id 8D0CC8819C for ; Tue, 21 Jul 2020 03:40:23 +0000 (UTC) X-Virus-Scanned: amavisd-new at osuosl.org Received: from whitealder.osuosl.org ([127.0.0.1]) by localhost (.osuosl.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id BZZ7ityYyGoU for ; Tue, 21 Jul 2020 03:40:21 +0000 (UTC) X-Greylist: domain auto-whitelisted by SQLgrey-1.7.6 Received: from mail-40140.protonmail.ch (mail-40140.protonmail.ch [185.70.40.140]) by whitealder.osuosl.org (Postfix) with ESMTPS id DE9C88816A for ; Tue, 21 Jul 2020 03:40:20 +0000 (UTC) Date: Tue, 21 Jul 2020 03:40:07 +0000 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=protonmail.com; s=protonmail; t=1595302817; bh=jAU4WV8TUxzOWOmy6luSSRdvupJIyR9sFqsVX6ePzlc=; h=Date:To:From:Reply-To:Subject:From; b=dVjb6q3Scxkm+ElBkAu+jVb5vGaXIpt8jM+YkQLam9yjhUeibGL1JOOXMcvsTnby3 V5ixobwTi+m09mHvivRecqwaj6r1jif6YkUMErhkA9OlRlL6cHJzH0fdND3mDJi2im 7Zy4HysmU1NV7xVocJKaGCJzX5uJPa5sQBqHyLys= To: bitcoin-dev From: ZmnSCPxj Reply-To: ZmnSCPxj Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Subject: [bitcoin-dev] Implementing Investment Aggregation X-BeenThere: bitcoin-dev@lists.linuxfoundation.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: Bitcoin Protocol Discussion List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 21 Jul 2020 03:40:23 -0000 Introduction =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D In a capitalist economic system, it is allowed for an entity to lend money = out to another entity, as long as both agree upon the conditions of the loa= n: how long, how much interest, any collateral, etc. This is a simple extension of basic capitalist economic thinking: that the = owner of funds or other capital, is the one who should decide how to utiliz= e (or not utilize) that capital, including the decision to lend (or not len= d). It has been observed as well that groups of people may have relatively smal= l savings that they can afford to put into investment (i.e. loaning out for= an interest rate), but as the technological capabilities of our shared civ= ilization have expanded, the required capital to create new businesses or e= xpand existing ones have grown much larger than most single individuals can= invest in. Thus, coordinators that aggregate the savings of multiple individuals, and = then lend them out for interest to new or expanding businesses, have also a= risen, in order to take advantage of the larger return-on-investment of mor= e capital-intensive but high-technology businesses, capturing the long tail= of small investors. Traditionally, we call these coordinators "banks". However, this typically involves delegating the work of judging whether a b= usiness proposal is likely to give a return on investment, or not, to the c= oordinator itself. Further, the coordinator typically acts as a custodian of the funds, thus a= dding the risk of custodial default to the small-time investors in addition= to loan default. (In this view-point, central banks that provide fiscal insurance in case of= loan default by printing new money, are no different from custodial defaul= t, as they degrade the monetary base in doing so.) This writeup proposes the use of features that we expect to deploy at some = point in the future, to allow for a non-custodial coordinator of multiple s= mall investors. This is not a decentralized system, as there is a coordinator; however, as = the coordinator is non-custodial, and takes on the risk of default as well,= the risk is reduced relative to a centralized custodial solution. Note that custodiality is probably a much bigger risk than centralization, = and a centralized non-custodial probably has fewer risks than a decentraliz= ed custodial setup. In particular, a decentralized custodial setup can be emulated by a central= ized custodial setup using sockpuppets, and without any decent sybil protec= tion (which can be too expensive and price out investments by the long tail= of small investors, thus leading to centralization amongst a few large inv= estors anyway), is likely no better than a centralized custodial setup. Focusing on non-custodiality rather than decentralization may be a better o= ption in general. A group of small investors may very well elect a coordinator, and since eac= h investor remains in control of its funds until it is transferred to the l= endee, the coordinator has no special power beyond what it has as one of th= e small investors anyway, thus keeping decentralization in spirit if not in= form. Non-custodial Investment Aggregation =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D In principle, if a small investor finds a potentially-lucrative business th= at needs capital to start or expand its operation, and promises to return t= he loaned capital with interest later, then that small investor need not st= ore its money with anyone else: it could just deal with the business itself= directly. However, the small investor still needs to determine, for itself, whether t= he business is expected to be lucrative, and that the expected return on in= vestment is positive (i.e. the probability of non-default times (1 plus int= erest rate) is greater than 1, and the absolute probability of non-default = fits its risk profile). We will not attempt to fix this problem here, only the requirement (as with= the current banking system) to trust some bank **in addition to** trusting= the businesses that are taking on loans to start/expand their business. (again: not your keys not your coins applies, as always; investors are taki= ng on risk of default.) The coordinator need only do something as simple as find a sufficiently lar= ge set of entities that are willing to indicate their Bitcoin UTXOs as bein= g earmarked for investment in a particular business. The coordinator, upon finding such a set, can then create a transaction spe= nding those UTXOs and paying unilaterally to the business taking the loan. The business provides proof that the destination address is under its unila= teral control (so that investors know that they only need to trust that the= business itself will do everything in its power to succeed and pay back th= e loan, without having additional trust in the coordinator to hold their fu= nds in custody). Then the individual investors sign the transaction, releasing their funds t= o the business. However, the issue now arises: suppose the business succeeds and is able to= pay back its loan. How does the business pay back the loan? Thus, prior to the investors ever signing the loan-out transaction, they fi= rst prepare a loan-payback transaction. This loan-payback transaction spends from a multisignature of all the inves= tors, equal in value to the loan amount plus agreed-upon interest, and dist= ributes the money to each of the involved investors. Crucially, this loan-payback transaction is signed with a `SIGHASH_ANYPREVO= UT` signature. Now, in order for the business to pay back its loan, it only needs to gathe= r enough Bitcoins to pay back the loan, and pay back the exact amount to th= e multisignature address of the investors. Then, any of the investors can reclaim their funds, plus interest, by re-an= choring the loan-payback transaction to this transaction output and broadca= sting it. The coordinator, for its services, may extract a fee from the loan-payback = transaction that all the investors can agree to; thus, it takes on as well = the risk of default by the business (the coordinator exerts effort to locat= e investors and encourage them to invest, and would lose the fee paid for i= ts efforts if the business it is proposing as a good investment does not pa= y back), which seems appropriate if it also serves as a basic filter agains= t bad business investments. Finally, by working in Bitcoin, it cannot have a lender of last resort, and= thus must evaluate possible business investments as accurately as possible= (as default risks its fee earnings). (investors also need to consider the possibility that the purported "busine= ss" is really a sockpuppet of the coordinator; the investors should also ev= aluate this when considering whether to invest in the business or not, as p= art of risk of default.) (the above risk is mitigated somewhat if the investors identify the busines= s first, then elect a coordinator to handle all the "paperwork" (txes, tran= sporting signatures/PSBTs, etc.) by drawing lots.) Thus, ***if*** the business is actually able to pay back its loan, the coor= dinator is never in custodial possession of funds. Cross-business Aggregation =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D Nothing in the above setup really changes if the investors would prefer to = spread their risk by investing sub-sections of their savings into multiple = different businesses. This gives somewhat lower expected returns, but gives some protection again= st complete loss, allowing individual investors to adjust their risk exposu= re and their desired expected returns. The batch transaction that aggregates the allocated UTXOs of the investors = can pay out to multiple borrowing businesses. And each business can be given a loan-payback address, which is controlled = by the investors that extended their loans. Investors generate an aggregate loan-payback transaction and signature for = each business they invest in. Collateralized Loans =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D As observed in https://lists.linuxfoundation.org/pipermail/bitcoin-dev/2020= -July/018053.html, a Cryptographic Relay would allow collateralized loans. Nothing prevents the "loan shark" in the collateralized loan example from b= eing a MuSig of multiple small investors. Practically, a coordinator would help facilitate construction of the necess= ary transactions and interaction with the loanee, but as long as ownership = remains controlled by the individual investors, there should not be any cus= todial issues. Of course, if the loan defaults, then the collateral needs to be sold in or= der to recoup the loss incurred in loan default case. Coordinating this sale amongst the multiple small investors is now potentia= lly harder. An additional service may be willing to pre-allocate Bitcoin funds into a t= imelocked contract, where the amount can be claimed conditional on transfer= of the ownership of the collateral to the service in the future, or if the= fund is not so claimed, to be returned to the service with the collateral = not claimed (as it might have been reclaimed by the loaner after successful= ly paying back its loan). This additional service earns by arbitraging the time preference: in case o= f default, the investors would prefer to recoup their financial losses quic= kly, while the service is now in possession of the collateral that it can r= esell later at a higher rate. Note that these are all operations that traditional banks perform; again, t= his idea simply removes the necessity for custodial holding of funds, in th= e way traditional banks do.