A Smart Contract Now Pays You to Borrow, is This Madness or is it Genius? – Trustnodes

A Smart Contract Now Pays You to Borrow, is This Madness or is it Genius?


The rise of decentralized and natively digital finance comes with a puzzle. Why are people borrowing dai at an interest rate of 20% from Compound’s smart contract, when they can collateralize it at 2.5%?

From their own supplied figures, it looks like people have offered 1.5 million dai and have borrowed about 1.4 million. You now need to pay 20% in interest to get some of that $100,000 worth of dai.

Dai disbalance at compound, December 2018.

Why anyone would pay that interest rate is not clear at all. Nor is it clear why arbitragers have not sent the interest rate close to that of collateralized dai at 2.5%.

Lack of liquidity might be one reason, with the ability to borrow dai on Compound a very new thing, launched only in the last few weeks.

Dai borrowing at 20% on Compound, December 2018.

But might there be something more fundamental? One possibility might be that there’s too much demand to borrow dai. In which case, the reason would be obvious too: it’s stable.

As one dai is one dollar and it stays so for months and perhaps years, you’d effectively be borrowing money without any paper work or anything and without the risk of the natural volatility of stablecoins.

If you borrow 1,000 dai today on Compound, that’s how much you owe tomorrow and the day after, plus interest. If you instead collateralize it through dai’s smart contract and eth’s price falls, then you get called. Meaning you don’t know, to the same extent, just how much you owe back if you utilize collateralized eth.

Easy collateralization of dai, December 2018.

It is difficult to see, however, why price stability would court such a huge premium of 10x. Obviously the lack of added risk that you might get called is valuable, but you can borrow fiat dollars for less than 20%.

The interest rate on Compound is set algorithmically depending on how much dai is supplied and how much is demanded for borrowing.

The aim of that algorithm is to ensure that if you don’t pay it back, then it can be taken in full from the collateral.

The system is somewhat complex because to borrow dai you need to put down 1.5x collateral of dai itself or any of the other supported tokens, including wrapped eth (weth).

That collateral you put down itself gets interest and not on just the amount borrowed by others, but on all of the collateral.

At a first look, this sounds a bit like a house of card and it may well be. Perhaps it is the tiniest bit of demand that has led to this stupendous interest rate for dai. Perhaps this isn’t being arbitraged because there might be something fundamental that makes the arbing not profitable.

Or maybe the algorithm is being played. There were suggestions that a whale was collateralizing (CPD) on Maker’s smart contract, sending the dai to Compound, then borrowing dai, sending eth to Maker on a loop of sorts.

Now if he is borrowing from himself, the borrowing rate is higher than the interest rate, so it makes no sense. On the other hand, if he is kind of cornering the market, he can set himself the interest rate and some milky bypasser borrows from him at what might be an inflated interest rate.

Obviously one can’t discount that this is Compound itself engaging in a marketing gimick, but we have no evidence of it so we are not suggesting it.

The other assets however, as shown in the featured image, have a near zero interest rate. The borrowing demand there, at least at face value, is far lower at just 1,000 weth for example, now worth $100,000.

Yet dai is of a different nature especially when it comes to borrowing. So maybe this is nothing nefarious but a very clever man or woman actually engaging in arbitrage, or maybe more correctly speculation, but of a different kind.

If you can borrow dollars at 20%, that might sound high but with the efficiency of cryptos and their quick and easy movements, you’re effectively leveraging yourself without the risk of getting called or having monthly payback demands.

Making this sort of a long term loan that doesn’t quite have to be paid back as there’s the collateral that can pay the 20%. So if you think an investment can give you more than 20%, you can just bet on it with extra money.

Now we say 20%, and maybe this is the trick. As you recall, collaterals are paid interest. So you put down 1.5 dais, and on that you get an interest rate of 16%. You borrow 1 dai, and on that you pay an interest rate of 20%.

Ah, it’s a flaw is it? 1.5 x 0.16 is 0.24. So he is paid 0.24. 1 x 0.2 is 0.2, obviously. So he has to pay back 20 pence while he is paid 24 pence. So he is paid interest of 4 pence?

Ma obviously he is paid interest, he is paid 24 on the collateral. Why only receive 4? Well obviously because now he has new money and he is paid to borr… what?

Well say how the math is wrong. We terrible at it, yes, but this is a bit simple. Twastnodes has 1,500 dai. We put that down as available for borrowing. We now getting an interest rate of 16% or 240 dollars. We can now borrow 1,000 dollars. This is new money. We still have 1,500. So now we have 2,500.

On this 1,000 dollars or dai we have to pay back 20%  or $200. So not only do we get an extra 1,000 dollars, but we’re paid for the privilege $40.

How on earth does that make sense? Who is paying us the $40? Well, remember we lent 1,500. We received 1,000. There’s still that 500 we have lent and someone has borrowed it and that someone is paying us about $100 at an interest rate of 20%.

Does this make any sense? What if they don’t want to borrow that $500. Well then the interest rate would be lower. The if and then algo keeps all in check and it so happens that sometime you don’t pay to borrow. You get paid to borrow.

Madness witchery. But is it true? Does it actually work? It kind of makes sense, but its just not how things are meant to work. Them young bankers and their new tricks making Hayek proud, perhaps.

Obviously this is a very new project and in this space we should expect things to go wrong, until proven otherwise, but in a way, logically, it kind of makes sense.

You’re lending more than you’re borrowing, so you get paid for that extra lending. Naturally demand is the other part of the equation. If people don’t want to borrow the extra $500, then the interest rate would be lower.

Robert Leshner, a Chartered Financial Analyst and founder of Compound, told Trustnodes that you’d get more by just lending the $500. That would give you $80 back, instead of you getting $40.

Obviously we put it to him that we have an extra $1,000 that we can use for whatever and then eventually pay back, plus $40. He says “you’re only losing if you do that.”

The front-facing web interface actually prevents you from putting down 1,500 dai and then borrowing 1,000 dai Leshner says.

Remember, if this does work as we conceptually understand it – minus any black-swan or a flaw – you get the 1,500 back through algorithmic “guarantees.”

Can we do that manually, we asked, the putting down of 1,500 dai to borrow 1,000, with the answer being in the affirmative in addition to “any sophisticated trader would see that it’s bad to do.”

Logically, this doesn’t quite make sense. The borrower is being paid to borrow. You can see how he is done so for that 500, but whoever borrows that 500 has to put down 750, so then we go to that 250 and so on turtles all the way down.

We must also say that initially we found Leshner to be somewhat evasive. We usually go speak to the innovators undercover, to get the “feels,” so maybe it is a bit unfair of us to say so, but maybe we kind of have to so that our readers exercise caution because again this is a very new project.

We can be wrong and we have been wrong, most recently in the BSV nonsense when very wrongly we gave it far more airtime than deserved as hindsight now shows. It is humans here, not laluminatus. So we make mistakes and perhaps proudly.

It may well be that perhaps there is a trick which pays you to borrow, but if we are to be honest, this sounds like usury.

There is a reason why that word raises emotions. Someone has to fall in the usury game of chairs. In times gone past, they used to forgive debts every 25 years or so. Now we have on demand filing of bankruptcy, which is pretty much the same thing.

But as self appointed judges without full information we must consider whether this is perhaps a trick more of the nature of magical code, rather than revolving wheels. And even if it is the latter, is it not somewhat amazing that usury might have been replaced by code.

The game of chairs, now on the world computer. If it is a game of chairs. Perhaps this is just genius, or just what it appears but if the latter then we’ll find out soon enough.

Because it is difficult to see how this can work when the borrower is being paid. Yet we have reasonable doubt because one can also see how it can work.

Remember that difference between $40 and $100. If you just lend $500, you get $100 at an interest rate of 20%. If you lend $1,500 and borrow $1,000, you get $40. There’s a $60 difference here. $20 of it has gone towards fees, so if you lend $500 you’d actually get $80, but $40 has gone where?

Now if we take all those $40s and we put them all together, maybe it works. Not for us to say. We can barely do 1+1 let alone analyze complex open source algorithms or engage in audits of code.

We iqra. Our job is to say hey this thing is here and they claim thus. Then others have to take it up and say whether this is madness or utterly genius.

Copyrights Trustnodes.com


Comments (5)

  1. great post! where is the attribution to the author? i feel its important to know who is speaking.

  2. After thinking about it a while, I’ve come to the conclusion that after the $1500 deposit, the $1000 loan is not new money. Its just a complicated way to hold your own money just as before the deposit was made. You can tell this is the case because the interest being earned is only on the remaining $500 so its the same as a deposit of $500.

    Also, could some kind of author attribution be added to trustnode articles? Thanks.

    1. Hi, We don’t name our authors so that they can speak freely, a bit like the Economist.

  3. How come the author doesn’t take into account the MakerDao stability fee of 18.5% /yr.? Obviously that makes opening up a CDP to generate Dai much less attractive, so (unless I am missing something) borrowing Dai from Compound at a rate of 20% is almost comparable to generating Dai from Maker’s CDP.

    1. This article is old. At the time, Dai’s fee was almost 0.

Leave a Reply

Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>