This Bitfinex Credit Bubble cannot end well

[N12]

The price on Bitfinex peaked on 4th December at 1175 USD (BitcoinWisdom). On that day, USD loans on the site were 6.8 million (bfxdata.com)

Ever since, the USD loans have ballooned up to ever increasing alltime highs while the price corrected, building an immense negative divergence that is still ongoing today: We now have 26.4 million USD in loans with a price of ~575. So the ratio once used to be 5787, and today it is 45913, constituting is a near nine-fold increase. It is true that Bitfinex has gained market share since then, but not by this big a factor, as is evidenced by only up to a ~2x increase in BTC loans as well as comparative exchange data over the past 6 months: http://data.bitcoinity.org/markets/volume/6m/USD?c=e&t=a&volume_unit=btc

Visualization (for some reason it does not include the actual price top that happened in December):



And here is what happened the last time around on Bitfinex when things went kaboom in a liquidation cascade (1 min intraday chart from Feb 10):

http://bitcoincharts.com/charts/bitfinexUSD#rg180zig1-minzczsg2014-02-10zeg2014-02-10ztgSzm1g200zm2g25zxzi1gVolzl

What on earth are these overleveraged maniacs doing? Are they waiting for the bubble messiah and not willing to deleverage, no matter what happens?!

[1715235061]

1715235061

[1715235061]

1715235061

[1715235061]

1715235061

[MoreFun]

They will rollback the trades and everything will be just fine, no need to worry so much about cascade.

[N12]

They will rollback the trades and everything will be just fine, no need to worry so much about cascade.

This is only true if the price on other markets returns to the point prior to when the cascade went off. If there is a differential, then that portion of liquidated traders will lose.

[N12]

It did decrease substantially (by 50%) in February. It just seems that whenever that happens, people take the longs back up pretty quickly.

I don't know if Bitfinex's data is real; I'm just assuming it is (same as any other exchange's data) for now.

[Kupsi]

They will rollback the trades and everything will be just fine, no need to worry so much about cascade.

They did that on Feb 10?

[N12]

They will rollback the trades and everything will be just fine, no need to worry so much about cascade.

They did that on Feb 10?

Not sure how it went in detail, but it seems they did something along those lines. Here are a few quotes from the operators back then: https://bitcointalk.org/index.php?topic=229438.msg7522014#msg7522014

Edit:

BFX gained significant market share over the past few months. The older data may thus be slightly skewed.

Skewed it surely is in part, but skewed by a factor of near ten? At the same time, the BTC swaps on Bitfinex have not increased by more than about a factor of 2. Also consider at whose expense they gained market share. Was this at the expense of leveraged trading?

[Dalmar]

BFX gained significant market share over the past few months. The older data may thus be slightly skewed.

[fonzie]

Increasing total sums of active swaps without paying interest

"So how can a lender manipulate this?

The most obvious way is to simply take all the lower rate swaps from the other lenders, forcing traders to take the higher rate swaps. Then close the unused swaps. Ofcourse the problem with this is that you're still paying interest on those unused swaps and BFX implemented a mandatory 1 hour interest charge on closing unused swaps specifically to deal with this sort of manipulation.

Game over? not by a long shot.

You can still take out swaps from other lenders and close them by closing them in a used margin position. For example, take a margin position of 1 btc. This takes up (at current rates) $600 of swaps. Now take out an arbitrary number of swaps from other lenders, say $600000 worth. Now just close "swaps used in margin positions" on the "total return swaps" page corresponding to that 1 bitcoin on margin. This subtracts the $600 margin from the $600000 of swaps you've taken out so you're left with $599400 of unused swaps....

... now repeat another 999 times....

Congrats! you've cleared out $600000 of swap offers from the orderbook while only spending interest on the swap in your actual margin position ($600). the 1 hour interest penalty for closing swaps only applies to closing unused swaps directly, it does not apply to closing swaps used in margin position."


SWAP demand wall also can be faked

"Placing huge swap demand walls is not that hard if even if you're not a whale. An idiosyncrasy about BFX is that you can stack multiple swap demands as long as they're each smaller than your tradeable balance. So even if you have only 1 bitcoin and an effective tradeable balance of $1000, you can just place a swap demand of $1000 a thousand times to give a wall of one million dollars. Give it a try, it costs nothing just to test it out. "

http://www.reddit.com/r/BitcoinMarkets/comments/28twpj/manipulating_bfx_swaps_for_fun_and_profit_a_howto/

[N12]

fonzie, the way I understand is is that it is used to manipulate the interest rates, but not the total active swaps which is what we are looking at. No new USD loans are created.

Also, the Bitfinex rep said he fixed it: http://www.reddit.com/r/BitcoinMarkets/comments/28twpj/manipulating_bfx_swaps_for_fun_and_profit_a_howto/ciehls0

[fonzie]

fonzie, the way I understand is is that it is used to manipulate the interest rates, but not the total active swaps which is what we are looking at. No new USD loans are created.

Also, the Bitfinex rep said he fixed it: http://www.reddit.com/r/BitcoinMarkets/comments/28twpj/manipulating_bfx_swaps_for_fun_and_profit_a_howto/ciehls0

The purpose defintely was to increase interest rates, but if you can take out xxxxxxx USD swaps without paying interest you would as side effect also increase the total sums of swaps.
There is no clear statement from him, in the reddit posts, that he has eliminated all of these "bugs/holes". The last time i tried it, 3 days ago, which was after his post, it was still possible.
All i can say is that i highly doubt that the actual data is completly "real" aka real swaps taken out from real traders used it in long positions.
Their trading code has/had a lot of  bugs, it was only 3-4 months ago when it was possible to build complete fake walls in margin trading up to 100000k BTC and you only needed 1 BTC to do so.
I played with that "feature" for a few weeks, and it took them quite long to realize that it was there, and it only got fixed after several users heavily complained about, even though people talked about it from time to time in their thread and also in the wall observer. I and many other have often seen 10-20k+ BTC walls in their orderbook for weeks, why didn´t they realize it???

TL:DR

Don´t give the stats to much credibility
You don´t know which other bugs there are that haven´t been discovered so far.

[tooil]

26 millions active loan is a fairly low number given the volume and market cap of bitcoin.

The maximum leverage ratio on bitfinex is 2.5, a somewhat reasonable number.

Active order is somewhat thin, this is the only element that worry me. As flash crash can happen if there is enough traders maxing their leverage and market happen to crash 40-50%. That will set off cascading margin call and they may not have enough market depth to absorb the volume.

[wilfried]

looking at the total sum of active swaps, the volatility of the total sum is ca $ 100.000,00 per day,
so ca $ 25.x mio are constantly long, thats awkward.

lets assume bitfinex traders that where there before feb 014 stopped trading because of non-predictable market conditions and started to lend out their money, it means we have new traders that are doing nothing but holding long positions. is that probable? no, i dont think so. imo its just a few traders holding long positions that make a huge sum and the many new traders that sum up to some $100,000.00 of volume, so if the whales liquidate, bye bye.. what do you think?

or maybe, there are some whales speculating on the market constantly going down and up, so they take huge long positions, wait until there is positive p/l, sell a little and if the price goes down again, they reinvest some money but only in small chunks, which enables them keeping the swap rate low by placing low swap demands..and what if those whales would get "special conditions" from bitfinex, for example "flash crash insurance", meaning market-rollback in case of margin call....?? then trading would be a save bet..or what if bitfinex itself went long with $24 mio - they are insured - they can roll back any time, they would make big money from the costs of interes of swaps and promote their platform at the same time..think about it. to ensure that doesnt happen, we would need transparency, what wont happen, we would need a system that indetifies each trade with a bitcoin wallet or so.

so what lenders should try: set a relative high rate (=>0.4%) and low swap period (2 days).
the system allows borrowers to cancel their swaps at any time, even if the money is in a position, so its somewhat clear, that new swap money leads to lower interest rates - new users are dumb (+/or inexperienced) and use frr or place too low swap offers taking the whole swap market down. if new money in-stream stops, rates will automatically go up.

having the feature of "frr" means an incentive for traders (cheap swaps), so the plattform is really made for traders, not lenders.

it would be nice to have a borrow-history, to see which amounts are borrowed.

[Sukrim]

I thought the main issue/risk would be the global margin:

26 million USD in long positions on BTC and LTC, vs. how much is actually backing them.
The same for BTC and LTC, then see in which market there is the highest margin/greatest risk.

Anyways, the most unexplainable thing for me is that they silently stopped using Bitstamp as second order book. They need every bit of market depth they can get, yet they claim to be "big enough" on their own now to handle stuff themselves? I understand that they started to cut MtGox off after they became less and less reliable with sending money - but they didn't even communicate Bitstamp being removed from their books...

[elebit]

Perhaps you are looking for something more sinister here, but from that graph it seems to me that the likeliest explanation of the ballooning swaps on Finex is simply that they have gained a lot of users since December. I don't know for how long they have existed but I didn't actually try trading there until last winter, as part of me moving away from MtGox, and I suspect that may be true for a lot of other users as well. I think you would find that everything from Kraken to Huobi gained a lot of users during that time period.

I disagree that USD swaps balloons more than BTC swaps in general in your graph. If you cut out June the trend is not nearly as pronounced. In fact, both swap pairs follow an upwards trajectory with two major exceptions: people were in general very bearish in February (omg! look at those slopes!), and people are very bullish in June. You can also see the formation of a bullish trend in January which later falls apart spectacularly, and I think we all know why that was.

These things are quite easy to spot in retrospect of course. The question is what it means for us right now, and the heavy slant towards USD swaps right now probably means that a lot of people are long on Bitcoin at the moment. That may lead to a spectacular dip if/when they get squeezed, but it could just as easily lead to a rallly. Everything depends on how the price develops during the following weeks/month. Exciting times ahead for all you fellow adrenaline junkies .

[thezerg]

Can you take a swap out at .1% (say) and then offer that same USD for .2%?  That would double the "total swaps" number without actually doubling the leveraged BTC. 

Also, I think its possible that people were taking USD swaps without purchasing BTC in an effort to encourage people to sell their BTC before the auction.

[zimmah]

Can you take a swap out at .1% (say) and then offer that same USD for .2%?  That would double the "total swaps" number without actually doubling the leveraged BTC. 

Also, I think its possible that people were taking USD swaps without purchasing BTC in an effort to encourage people to sell their BTC before the auction.

i don;t know much about swaps but i believe this may be the case here

[Sukrim]

Can you take a swap out at .1% (say) and then offer that same USD for .2%?  That would double the "total swaps" number without actually doubling the leveraged BTC. 

No, you can't. Where would you pay interests ("swap payments") from?! You can't guarantee that your 0.2% offer would be taken to cover the interest.

The time frame of this chart is also so large that an auction of just 30k BTC is not really a big factor in there.

[thezerg]

Can you take a swap out at .1% (say) and then offer that same USD for .2%?  That would double the "total swaps" number without actually doubling the leveraged BTC. 

No, you can't. Where would you pay interests ("swap payments") from?! You can't guarantee that your 0.2% offer would be taken to cover the interest.

The time frame of this chart is also so large that an auction of just 30k BTC is not really a big factor in there.

I wasn't asking whether it was smart to do so, I was asking whether it is possible.  WRT is it smart: personally, I've seen cases on the BTC side where it was worth the risk of not being able to resell it.  It happens when somebody's "automatic re-lend" at .005% pops up yet the average rate for days had been around .02+.

[gizmoh]

Can you take a swap out at .1% (say) and then offer that same USD for .2%?  That would double the "total swaps" number without actually doubling the leveraged BTC. 

Its NOT possible.

[urwhatuknow]

Guys

Since when so called "moderators" are starting FUD threads?

Cui prodest?

Ask yourself this simple question and if your IQ is not lower than your shoe size you'll come up with the right conclusion.

Have a good day and a great weekend

Giancarlo
Bitfinex Team