Finance : no more middlemen ?

Can you explain the concept of decen­tra­li­sed finance, or “DeFi” as it is now called ?

Decen­tra­li­sed finance pro­vides access to finan­cial ser­vices such as cre­dit or asset exchanges (main­ly cryp­to-cur­ren­cies). Unlike tra­di­tio­nal finance, there are no ban­king or legal inter­me­dia­ries. Exchanges are mana­ged bet­ween pri­vate com­pu­ters thanks to the blo­ck­chain, which makes it pos­sible to gene­rate contracts that are self-exe­cu­ting under cer­tain condi­tions (e.g. repaying a loan at regu­lar inter­vals or com­mit­ting to sell an asset at a given price). This is what we call “smart contracts”. They replace banks for tran­sac­tions and the legal arse­nal (regu­la­tors, lawyers, courts) that was pre­vious­ly cal­led upon in case of breach of contract. The blo­ck­chain is neu­tral, unal­te­rable and allows to trace in all trans­pa­ren­cy all the steps of a contract. It is an open book sha­red by all.

A plat­form like Unis­wap records an ave­rage of more than $1bn-worth of tran­sac­tions per day¹.  Does this mean decen­tra­li­sed finance is taking off ?

Com­pa­red to tra­di­tio­nal finance, it’s tiny, but these volumes are mas­sive and conti­nue to grow. So, we can indeed talk about a take-off, espe­cial­ly for len­ding pro­to­cols (Com­pound, Aave), decen­tra­li­sed exchanges (Unis­wap, Curve) and “sta­ble­coins” (Maker). This suc­cess is part­ly dri­ven by the explo­sion in capi­ta­li­sa­tion of cryp­to-cur­ren­cies, which is still close to $2tn².

But isn’t decen­tra­li­sed finance only for cryp­to-cur­ren­cy holders ?

Yes, only cryp­to-cur­ren­cies can be tra­ded. Exchanges of regu­la­ted assets, such as stocks, are not pos­sible for rea­sons of non-com­pliance with the legis­la­tive fra­me­work. It is impor­tant to unders­tand that the clas­sic mar­ket works in reverse to the cryp­to-cur­ren­cy mar­ket. In mar­ket finance, we know the iden­ti­ty of the players (thanks to “know your cus­to­mer” pro­ce­dures for iden­ti­fying cus­to­mers), yet their ope­ra­tions some­times remain opaque. In decen­tra­li­sed finance, iden­ti­ties can be pseu­do­ny­mi­sed, but all ope­ra­tions are transparent. 

What does “DeFi” pro­mise to achieve ?

To take an extreme case, the ini­tial ver­sion of Unis­wap was only 300 lines of code. Hence, decen­tra­li­sed finance is acces­sible to the masses as oppo­sed to the costs of ente­ring tra­di­tio­nal finance, which are very high (licen­sing, ini­tial capi­tal, lawyers). If these costs come down, it will sti­mu­late com­pe­ti­tion and access to com­plex finan­cial pro­ducts that are today reser­ved for the richest. This demo­cra­ti­sa­tion is the strong pro­mise of DeFi. That being said, it is not without dan­ger. Most regu­la­tors believe that one should not give too easy access to finance because small hol­ders could risk losing out if they lack full unders­tan­ding ; this is what we some­times see with hyper-sim­pli­fied stock tra­ding mobile apps like Robinhood.

What are the obs­tacles to large-scale adop­tion of “DeFi”?

In addi­tion to the regu­la­to­ry hurdle men­tio­ned above, the indus­try also has to over­come tech­no­lo­gi­cal bar­riers. The main one is sca­ling up to a much lar­ger scale. Des­pite the rela­ti­ve­ly mode­rate num­ber of exchanges, there is alrea­dy some conges­tion on the Ethe­reum blo­ck­chain. There are often too many tran­sac­tion requests com­pa­red to the vali­da­tion capa­ci­ty of Ethe­reum. Fur­ther­more, mining is struc­tu­ral­ly slow com­pa­red to the thou­sandths of a second requi­red for an exchange in cen­tra­li­sed finance. An inter­ope­rable net­work of mul­tiple blo­ck­chains and side­chains will like­ly be requi­red to deve­lop DeFi. This is a signi­fi­cant tech­no­lo­gi­cal chal­lenge, but many deve­lo­pers are wor­king on it and given the dyna­mism of the sec­tor their inge­nui­ty should not be underestimated.

Doesn’t decen­tra­li­sed finance encou­rage ris­ky speculation ?

As finance does in gene­ral, I would say ! That being said, there is a spe­cu­la­tive fren­zy around cryp­to-cur­ren­cies. If we look fur­ther ahead, we can hope that this spe­cu­la­tion will fall back to make way for solid mar­kets based on blo­ck­chain tech­no­lo­gy, with an inter­me­dia­tion cost limi­ted to the remu­ne­ra­tion of the net­work infrastructure. 

What is the risk of a cryp­to­cur­ren­cy mar­ket reversal ?

It should not be over­loo­ked. In a context of high valua­tion of finan­cial assets, we can indeed fear a violent rever­sal of the cryp­to-cur­ren­cy mar­ket. What’s more, there are doubts about the resi­lience of sta­ble­coins³ to such a sce­na­rio. It is easy to ima­gine a ban­king panic simi­lar to the one suf­fe­red by money mar­ket funds during the 2008 cri­sis ; except this time cen­tral banks will pro­ba­bly refuse to save sta­ble­coins by exer­ci­sing their power of len­der of last resort. Hence, it is beco­ming urgent to build tools to anti­ci­pate sys­te­mic risks in DeFi, a task we are acti­ve­ly wor­king on in our research chair.

Can the high ener­gy consump­tion of blo­ck­chain slow down – if not prevent – its development ?

As we face the acce­le­ra­tion of glo­bal war­ming, we can­not ima­gine DeFi relying on Bitcoin’s proof of work and its exces­sive car­bon foot­print – which cur­rent­ly accounts for 0.3%-0.5% of the world’s elec­tri­ci­ty consump­tion. The future will pro­ba­bly be found in proof-of-stake, which uses much less ener­gy. In proof-of-work, the miners in charge of upda­ting the blo­ck­chain must solve an ener­gy-inten­sive cryp­to­gra­phic pro­blem. Whe­reas proof-of-stake, ins­tead of selec­ting miners based on their amount of work, the pro­to­col uses the amount of cryp­to-cur­ren­cy they hold. The idea is that miners who hold cryp­to-cur­ren­cies native to the Blo­ck­chain also have an inter­est in it wor­king properly.

Proof of stake has alrea­dy been tes­ted, in par­ti­cu­lar by Tezos, a blo­ck­chain co-foun­ded by a poly­tech­ni­cian, Arthur Breit­man, and whose research cen­ter Noma­dic Labs is one of the spon­sors of our chair. We are conduc­ting work to for­mal­ly esta­blish and, as far as pos­sible, improve the robust­ness of the proof of stake.

Interview by Clément Boulle