Keeping up with Crypto from a Recordkeeping Perspective

Keeping up with Crypto from a Recordkeeping Perspective
07/12/2022

This article covers recordkeeping, the less glamorous—but very necessary—part of the new digital asset class of crypto. It examines the technology needed to keep up with crypto from the unique perspective of the CEO (a fund-administrator-turned-technology-provider) of the Titan Platform by AdvancedAIS. It's intended for fund managers doing their own back office, fund administrators, and anyone curious as to what the challenges are with crypto and why not all fund administrators/software platforms offer crypto functionality. Let’s dive in.

What Prevents Legacy Systems from Keeping up with Crypto? Volume, Volume, Volume!

Volume Challenge #1: Increase in Daily Trades

Legacy systems were originally built to support 30,000 trades per day. From a fund accounting perspective, this means that after a manager makes a trade, it is duplicated in a portfolio accounting system, costs are tracked, relevant foreign exchange (FX) rates are factored in, and journal entries are generated. It gets more complicated if there is a sell or cover, as the system then needs to determine what cost methodology is being used—FIFO (first in, first out), LIFO (last in, first out), HiCost, LoCost, Specific, HiTax, LoTax, etc.—and apply this methodology. So if you’re selling Apple using FIFO, the system needs to go through all the portfolio holdings at the tax lot level, and then figure out which lot(s) to relieve the sale against. Finding the first available tax lot in 30,000 daily trades to relieve the sale against isn’t a problem. Three million trades, however, can be challenging. And that’s the volume we’re seeing in some crypto strategies.

Volume Challenge #2: After-Market Trades

Supporting three million daily trades, including sales that need to be matched up to tax lots using FIFO, is time-consuming enough. An extra challenge comes in the form of additional trades that are out of the market. They still need to be factored in, so systems need to be able to recalculate cost relief on millions of trades at the touch of a button.

Volume Challenge #3: Cross-Exchange Trades

Arbitrage across exchanges is the next domino in the volume challenge. Managers are trading at the microsecond level across exchanges. But these exchanges don’t talk to one another or have standardized data or reporting. So, for example, when these millions of daily trades are downloaded, you may see 20,000 Bitcoin disappear from one exchange. You may then see 19,999.99999985 units appear in another exchange in the same microsecond. When this happens, you can match the trades up and assume there’s a transfer and the difference is a fee. This process is easy enough to go through for 10 or 15 transfers, but the volumes are usually much too high to accommodate a manual process.

The solution is to employ a platform that can amalgamate all activity, look at microsecond-level time stamps, use artificial intelligence to determine transfers and fees, and build the chronological trade order from exchange transfers, decentralized exchange (DEX) trades, and over-the-counter (OTC) movements.

Volume Challenge #4: Reporting

A year ago, one of our fund admin clients was working with an audit firm that asked for an exhaustive Excel list of a crypto fund’s transactions. Excel includes a maximum of 1.9 million lines, so for many crypto funds, a single file won’t capture a full year’s worth of crypto transactions. And you wouldn’t want it to—the file would be huge! When the auditors continued to insist, the fund administrator emailed them several files of trades which were so large that Excel froze every time the auditors tried to run their macro-based sample testing programs. Service providers need to be able to think beyond their traditional approaches. In this case, Titan provided JSON files to accommodate reporting volume.

What Are the Consequences of Not Having This Information Available?

The 2021 tax season highlighted many shortcomings occurring in the crypto landscape. Some investment managers were in hot water with the tax authorities because they couldn’t provide a realized gain/loss schedule. Certain administrators had to explain to the auditors why they didn’t have readily available transaction details.

The consequences are wider spread. If you can’t track it, you can’t audit it. And if you can’t audit it, you can’t regulate it. And this can lead to knee-jerk reactions from governments.

Where Is the Tech to Keep up with the Tech?

If a fund accounting software can track long/short equity funds, why can’t it handle crypto? The underlying problem is that most fund accounting platforms are 25–30 years old and were not designed to adapt. Many legacy systems simply weren’t built for the kind of volume we’re seeing in the crypto space. Some were designed before the cloud existed, so the speed and the adaptability just aren’t there. When trying to account for crypto, legacy systems can get tripped up by some additional factors, including decimal places, in-kind subscriptions, hybrid funds, and exchange communication.

  • Decimal Places: The majority of systems were designed to account for two decimal places in pricing. Maybe even four if they’re somewhat forward-thinking. Fair enough. Who could have envisioned an asset class whose assets are priced out up to 45 decimal places?
  • In-Kind Crypto Subscriptions: Most systems have been designed to accommodate, for example, a GBP class within a USD class fund. Increasingly, however, investors are subscribing in-kind to a USD fund with Bitcoin and want their share class denominated in Bitcoin.
  • Hybrid Funds: Fund technology platforms have historically been designed to support either hedge funds or private equity (PE) funds. As the lines continue to blur between these types of funds, historical systems aren’t designed to accommodate the characteristics of each. For example, a PE fund with a liquid portfolio of Bitcoin and Ethereum often ends up being tracked on a different system or (gulp) on an Excel spreadsheet.
  • Communicating with Exchanges: In the beginning, exchanges didn’t follow a standard when setting up their data flow. Each exchange was set up differently. Since then, the process of standardizing data has improved, but it’s still a bit painful. Unlike with traditional brokers, administrators and investment managers don’t receive data from an exchange via a CSV download. These parties need IT resources to help bring data in, because data these days is transferred via an authenticated API, which is often beyond the skillset of a typical accountant (this one included!).

Here’s a funny story to demonstrate how accountants are always the forgotten ones. In the early days, some exchanges were set up with only one kind of access, with which managers could buy and sell crypto and access their records. Clearly, the managers wouldn’t and shouldn’t have given access to anyone else but themselves. So until the exchanges caught on that they also needed to provide read-only access, the investment managers had to download the reports daily and send them to their fund administrators.

At the time this article is being written, some exchanges will send information only in response to very specific requests. For example, users can ask for records of transactions that involve a sale of BTC for ETH, but they can’t request a listing of ALL trades. Many combinations of trades and new securities are being created all the time, so this is not an efficient process to gather information. It also isn’t a process that auditors can rely on to satisfy their completeness requirements.

Where Are We Now?

There is a big demand for fund administrators that can keep up with crypto. But not all fund administrators have systems that can support this asset class. There are some “add on” programs that can bring in the trades, but having a non-integrated program at such high volumes increases data transfer risk.

Some administrators have decided that they won’t administer crypto funds, for whatever reason—be it lack of technology or incompatibility with their risk appetite. But there are also investment managers who dive into this asset class, and their administrators find this out after the fact. They’re then left scrambling to figure out what to do.

Some administrators are taking on crypto funds with their own proprietary systems that haven’t been able to keep up and have let their investment manager clients down. This has a “once burnt, twice shy” effect with these investment managers who are not just switching administrators but looking to license an in-house system. (I haven’t marketed heavily to investment managers and was surprised to receive five calls in one week from various managers who all wanted to do their own back office.)

We’ve heard rumors of some companies using NAV Lite (simply using their ending balances as their net asset value, or NAV value). But this kind of approach completely ignores relief methodologies and doesn’t discharge a fund’s duties of keeping adequate books and records.

Will the Legacy Programs Catch up to Crypto?

They might, if they’re designed to adapt to high volume, to be flexible enough to change, and to test and deploy changes efficiently.

Modern programs employ a technique called Test Driven Development, or TDD. In this methodology, a test is written at the beginning of programming a new functionality. Something like, “If we buy a Canadian option in our USD denominated fund for $A with a fee of $B at an exchange rate of $C and it sells for $D at an exchange rate of $E in Canadian dollars, we expect to see a realized gain of $F and here are the journal entries we would expect to see generated.” That test is saved, along with hundreds of thousands of other tests, as the program is built. When something new (e.g., crypto transfers) is programmed in, these hundreds of thousands of tests are run to ensure that nothing has broken or changed in the process.

This TDD methodology wasn’t created until about 20 years ago. This means that older programs do most of their testing manually, which of course makes it harder to keep up with releases. This manual testing also makes these systems more prone to error and much more expensive, as they require armies of manual testers.

What’s Next?

Daily trade volumes keep increasing. And the pace of change continues to increase.

There seems to be something new every two weeks in the crypto space. From a technological support point of view, keeping up with what is coming next down the pipeline—be that decentralized finance (DeFi), microsecond trading among exchanges, movement of data, etc.—is a constant.

No finish line is in sight when it comes to supporting the finance industry. There will always be changes and new demands—new types of swaps, more stringent privacy rules, regulatory changes that allow over 2,000 shareholders, or entirely new asset classes. To stay relevant, modern systems don’t just need to be able to handle change—they need to expect change and plan for it. Even if we don’t know what the changes are going to be, we can build the software programs to be able to pivot when needed.

The original risk concerns around crypto custody and ownership have shifted as the asset class has become better understood. The landscape continues to evolve and the administrator crypto challenges may look totally different in a year. But one thing holds true: these new challenges highlight the fact that modern and adaptable technology is needed to keep up with crypto.