Hacker News Reader

Comment by paulsutter | Hacker News Reader

Surprise - a good idea from the Winklevii. The convenience of buying and selling bitcoin from the comfort of an ordinary brokerage account, and making it possible to hedge bitcoin by selling and shorting.

Anyone know how to estimate whether there's enough liquidity in existing bitcoin exchages to support it, and allow them to track the price accurately in the ETF?

View on HN

> Anyone know how to estimate whether there's enough liquidity in existing bitcoin exchages to support it, and allow them to track the price accurately in the ETF?

Yes. The daily volume in the ETF must be a small fraction of the daily volume in the underlying asset, and it must be possible to both buy & short the underlying.

So net net: An ETF is incredibly premature. This is the Winklevii taking serious advantage of folks who don't understand ETFs in depth.

Edit: Clarified the relative quantity.

Yeah, the whole public BTC order book on MtGox can be had right now for under $2mm USD. The other exchanges are smaller yet.

The market is very loud, but still very tiny. This will change, and it's good to start early, but I agree wholeheartedly that an ETF is premature.

This will not be the case in 12 months, though.

I think the biggest risk to a fund like this is data security. If you have more than 50k bitcoin sitting around, APT doesn't even begin to describe the measures people will take to steal your keys.

Their S1 specifically states that the fund is not holding the Sponsor company (holding the keys) liable should their proprietary Security System fail and the keys get hacked and the coins are stolen unless it was gross negligence.

This is going to be a bumpy ride for those who don't have a deep understanding of security surrounding crypto. Based on some of the technical mis-descriptions in this doc, I am tempted to speculate that they may not have the requisite understanding to fully model the threats they face. Then again, government paperwork takes months, and perhaps this is just an early step and they'll source all of that in the interim. (Hey, tall dudes: I've got 15 years of experience keeping private data away from super-determined hackers and other bad guys, and my email address is jp@eeqj.com! Get in touch!)

The filing says nothing about the details of their proprietary Security System.

Following that, the future of liquidity is a huge unknown void right now. I wouldn't be surprised if at some point in the next 12 months, there are no exchanges on which you could liquidate more than $50-$100k in bitcoin.

It's even difficult right now when everything is nominal - anyone who thinks buying a few million dollars in bitcoin is easy or simple has never even begun to try to do so.

"the whole public BTC order book on MtGox can be had right now for under $2mm USD"

No it cannot. Some people have placed sell orders at $1000, $10k, $100k, etc. To the point it would take trillion and trillion of dollars to buy up all BTC.

But you can be sure that if you tried to do this, a lot of other sell orders would pop up, making it effectively impossible to buy up all bitcoins.

> But you can be sure that if you tried to do this, a lot of other sell orders would pop up, making it effectively impossible to buy up all bitcoins.

Yes, at entirely reasonable prices, right? :D A substantially higher cost basis starts to cut into fund profits...

Also, a minor correction to your comment: the depth of the MtGox order book is public, and it was accurate at the time of my comment. It's up to $2.2mm USD to buy the whole thing now.

Does the order book show all the "dark pool" trades? It surely cannot show all those who sit and wait (or have bots) for the price to change so they can move their ask price up.

The order book would show people who are sitting on either side of the market, so if they have a bot that's adjusting their quote it would still show it - unless I'm misunderstanding your question?

I think we're talking about parties who don't actually have an order in. Rather they have automated a process to enter orders in response to certain conditions. Depending on what those conditions are, you won't be able to rely on a continuous price trajectory.

Admittedly, I don't see how this is different for BTC than it would be for any other asset.

Those are not providing liquidity in a useful sense right now. Usually the measure is people prepared to trade at current bid or offer ie a very narrow measure. If you want to buy at 1% down that's not useful now.

So... if I put a sell order for 1BTC at $10 million, their number would go up to $12.2M?

Whatever site you are using to see the order book does not show it in its entirety. It really is a lot deeper than $2.2M. I know for a fact because there has always been sell orders at ridiculous amounts such as 1 BTC at 1 trillion dollar.

For example bitcoincharts.com now truncates the order book (2 or 3 months ago it was showing everything, on the sell side at least).

> Following that, liquidity.

There needs to be an effective method of shorting in order for the ETF to be effective. I think it will take much more than a year.

I don't follow that logic, could you explain it?

https://news.ycombinator.com/item?id=5975704

A thought occurs to me: Don't the twins own something like $10M of BTC? Perhaps they personally will provide liquidity for this ETF.

Yup. They claimed they owned "1% of all BTC", without specifying if it was 1% of 11M (current amount) or 21M (max theoretical amount). But if it is 1% of the latter, then it matches perfectly with the number of bitcoins needed to back this ETF (1M shares * 0.2 BTC/share = 200k BTC).

That probably kills my idea, since to provide liquidity for the ETF they'd need to hold significant BTC outside the fund.

> * Perhaps they personally will provide liquidity for this ETF.*

That sounds suspiciously like "unloading".

I'm no lawyer, but I'd be surprised if it's legal for ETF providers to seed their own fund. The need for authorized participants is rather important.

I expect we'd see more liquidity on MtGox if they figured out how to write a matching engine. I would never add liquidity on a venue that routinely takes hours to process orders and cancels.

I have had a theory that most traditional matching engines can get away with lower latencies because they can multicast their marketdata.

It seems like sending out all marketdata events to each connected user over a tcp connection is bound to be problematic.

Perhaps they will add a collocation site where active Market Makers could receive multicasted marketdata with much lower latencies.

The price of ETF and the price of underlying are simply correlated as a result of market forces. There's no "tracking" between the two. The only "tracking" happens to the Net Asset Value of the ETF, which the quantitative fund manager must aim to follow the underlying as much as possible. In this case, by holding 100% of asset in Bitcoin in this trust, the NAV tracking can be assumed to be perfect.

The liquidity doesn't matter if the ETF has a fixed pool (with regular creation and redemption). The volume of the ETF can exceed the volumes of all Bitcoin exchanges combined, because the investors are actually trading paper - certificates that represents some units of the trust, which holds Bitcoin.

If creation needs to happen, the investor needs to give the underlying to the trust in exchange for units. In this case, only BTC changes hands.

If redemption needs to happen, the investor will receive the underlying from the trust. In this case, only BTC changes hands.

So, at the trust's side, it's entirely possible to eliminate all fiat transactions after the ETF is issued. All future changes in units will happen with the underlying only, which is Bitcoin. This is similar to and consistent with most of the stock indices ETFs out there in the market.

An analogy: The liquidity of gold ETFs can obviously exceed the liquidity of physical gold. Actually it already does (for retail investors).

Low transparency tends to lead to high transaction costs, low liquidity, and high volatility in a market -- which is pretty much what's going on with BTC right now. There's also very limited commercial exchange volume, from what anyone can tell, further hindering liquidity. Seems like it'd be pretty hard to earn a decent return from this ETF.

I'll be watching this one from the sidelines. Definitely not participating.

"Someone on one of Team Macro Man's chats recently described this ETF as a Rorschach test of what people think of digital currencies. That is utterly wrong: it's an IQ test. Bitcoin is anonymous, untaxed (for now) and quite liquid in and of its own right despite all the complexities of a cryptocurrency. A bitcoin ETF is taxed, has fees, may or may not be liquid at all. So, this is really a test: do you want to facilitate the exit of the Winkelvii from an investment at inflated levels which they will soon be taxed upon or do you want to sit this hand out? "

Macro-man gets it.

If the ETF grows in size and its net asset value must assume a large fraction of the outstanding bitcoin stash or towers over the daily trade volume, we could have some issues here...

I didn't read the SEC filling. But there are 2 major kinds of ETFs: those which trying to accurately track the index and those which tracking daily movements (usually leveraged). Many ETFs tracking commodities are of the second type, so buy&hold investors will lose money by buying them.

Reply on news.ycombinator.com