Data-Driven Crypto Economic Modelling - Philip Gradwell - Chainalysis

Data-Driven Crypto Economic Modelling - Philip Gradwell - Chainalysis

. 19 min read

About this episode

Learn how to say "Chainalysis" from Philip Gradwell, the company's chief economist and data driven analytic. He chats with host Samuel McCulloch about what on-chain economics is and their use in day to day operations and analysis.

Philip Gradwell is Chief Economist at Chainalysis, the blockchain analysis company. He analyses on-chain data to understand cryptocurrency markets. This includes identifying economic fundamentals, how cryptocurrency moves on-chain between exchanges and across borders, and the nature of crypto crime, amongst other topics. Prior to joining Chainalysis, Philip led a team of economic consultants working globally on energy system analysis and climate change economics.

Linkedin | Medium | Twitter

Chainalysis' Website

Where to find the show

iTunes | Spotify | Stitcher | Youtube | RSS Feed

What to listen for

  • Why blockchain has much less data precision than people think and why the datasets are noisy.
  • Why Chainalsysis’ core work is really about mapping datasets of entities and their relationships to one another.
  • Why the interpretation of the data in the Crypto-economy is so much more complex than in Fiat economic systems.
  • What the biggest challenges facing on-chain economic studies are and what Philip will be working towards solving in 2020.
  • Why the creation of Stablecoins is about the demand for a dollar-denominated product which is regulated in a different sphere of jurisdiction from the one the underlying asset originated from: the USA.
  • Why Tether was essentially filling a gap in the market for a stable fiat on-ramp and meeting the demand for a privately issued tokenised dollar in 2018.
  • Why Chainalysis’ on-chain metrics during the Bitmex implosion provided data-driven insights into the market structures and the price movements of Bitcoin.
  • How daily inflows and outflows of cryptocurrencies from exchanges correlate to price movements due to the illiquid nature of Bitcoin and cryptocurrency markets.
  • Why Philip believes the fundamentals of Bitcoin are stronger than any other Cryptocurrency but that DeFi still has many maturity cycles to go through.
  • Why we will be exploring a new paradigm in terms of monetary and fiscal policy in the fiat world after the Coronavirus induced recession.
  • Why Philip worries as an economist that we had not gone far enough into the ‘future’ to see the consequences of the last ten years’ expansionary monetary policies after the 2008 crisis, but we have just doubled down with the economic responses to the Coronavirus in 2020.


If you like End of the Chain you can help support the show by doing one of the following:


Ethereum: 0xDAb148614f22dDa800cF006Be7932eeEB75AC644

Show Notes

Chainalysis analyse chain data for regulatory analysis and for clients that need to trace transactions on the blockchain and how they are tied into the real world.

What does a Chief Economist do? Chief Economist is one of the stranger roles in an organisation like Chainalysis.  Because the core of the business is about understanding transactions on the blockchain and how they link to the real-world entities, this is mostly used by law enforcement agencies and for compliance purposes. If you look at a blockchain it is a complete economy, so its individuals sending money between themselves in exchange for services. I am here at Chainalsysis to try to understand the crypto economy from that perspective and to take the macro view and describe the trends we see in usage patterns in Bitcoin, Ethereum and all the cryptocurrencies we cover.

Chainalysis covers a lot; they work with some of the largest exchanges in Crypto. They also are providing services for a lot of people. Compliance is an integral part of every real-world businesses’ operations. Employing a company like Chainalysis is important to make sure no company taking funds is connected to criminal or terrorist financing. I have always been interested in Chainalysis because I come from an intelligence background. The gobs of metadata that exist and can be connected with Bitcoin is both extremely interesting, and it also scares me. It scares me in one respect. If you think about the level of data collection, and you can pair the analytics you have with the data, a government might have in their own scale intelligence-gathering operations, it becomes a whole different picture.  On the private side, when you are only providing compliance, then maybe it is a bit different.

Blockchains are databases. At the heart, they are database that says this address transacts with this address and actually the core thing we do is we map core sets of addresses to entities because any individual or business can have as many addresses as they want.  Entities are individuals and businesses are things like exchanges.  This entity looks like a business, and this one looks like an individual. This allows our clients to follow the funds in from these addresses for their services. This is basically the core of what we do. Then you mentioned the metadata which is on the blockchain, but it is quite limited in nature actually.  There is also the peer-to-peer network in which transactions are broadcast, but the way in which transactions are broadcast makes it less useful than you might think. The core data set is this mapping from entities to businesses as a real-world thing.

The choke point is really the service where criminal activity can be stopped. This is where governments focus most of their work.  The only powers that can be used are at the exchange level.

If I am a business and I am going most of my business on the blockchain, it is publically verifiable transactions that people could use these data flows against a company. I do think Bitcoin is too transparent.  People can see what I have – either we see greater privacy options in cryptocurrency or we will get more comfortable with this level of transparency.  One example of this is crypto reserves and exchanges showing clearly their cryptocurrency reserves to prove they have not run off with my Bitcoin. I think the exchanges have many reasons why they don’t want to do this. But their nervousness about revealing how big their business is.  I don’t like the proof of reserves as it is disingenuous.  For a business, as soon as you make a deposit and then lend out based on this deposit; you have created more money, in essence, and you don’t know what their loan book looks like.  I think it can give you some help in terms of ratios.

People think blockchain has much more precision than it actually has to be honest. I spent two years at Chainalysis doing fundamental research to make sure that when we say something, we can say it with certainty because we know why we may be wrong, why it may be a high or low bound. The data sets are noisier than people think. It’s not quite this God view of the crypto-economy. It takes a lot of work even to get to this point. Every day I ask myself the same question, how do I know I’m wrong?  Because most of the time, you are.   Again it mostly comes back to the service providers. With zero transparency into what was happening on these exchanges, there were attempts to show liquidity via volumes of trading.

Do you like that all this data is public? The data is fascinating from an economics perspective. The thing that is really interesting from the crypto-economics side; you mentioned earlier you’ve got the address data set and then you have to do all this work to link these to entities. So you get to look at a complete economy, but it is like you have blurry vision; because you are never sure if these addresses look like they should belong to this entity or if they actually belong to another entity.

I had a whiteboard my first week at Chainalysis listing on the metrics we wanted to do, and it has only taken eighteen months later to do all our research, which means we can finally do all these metric charts.  The interpretation of the data is so much more complex than economics in the fiat system. The entities are very clear cut in the fiat system.

What are the biggest challenges facing on-chain economic studies that are yet to be solved or that you are working on in 2020?  It is an interesting time for Crypto as we are recording this during the coronavirus lockdown and the really significant price changes we are seeing and whether crypto winter is coming or if Crypto can survive even. I think there is the broader narrative of what is Crypto, is it heading into Winter? Is it even going to survive? Are the people who were holding Bitcoin are they even bothered about the price move and did they move into exchanges? These are all tactical questions that get asked the most. Then there are bigger questions, such as what is the interest rate of Bitcoin? What’s the equivalent metric of GDP growth rate in Crypto? We still don’t have a framework to explain this asset and why it is useful.  It will often come down to some high blown narrative trying to convince people to get involved in the tech stock gamble.  But I want to give more data on actual use cases and real valuation methods, both on the theoretical side and also bring the data to it. That is the bigger challenge.

Where do you stand on the long term valuation and worth of  Bitcoin?  Yes, I am not a maximalist. Of the Economists who are in this space, I am not the typical economist; I see cryptocurrencies a really interesting alternative asset that has characteristics that others don’t. If it is a store of value, it can be used as means of exchange; unlike your bar of gold in the apocalypse. Bitcoin is, in many ways, a basket of very interesting properties that get fitted together. I think the world is short of assets that are not correlated to what is going on in a certain country, that have a fixed ownership policy but can be stored in very large amounts safely and securely.

Stablecoins, like Tether, removed the volatility from the Bitcoin model makes sense. The tokenisation of the dollar might change some of how the financial plumbing will be done.  We need to see how large the world’s demand for Bitcoin is in the long term. The best example of why people love Tether – People in Russia would send their money back to China using Tether instead of Bitcoin.  This asset was dollar-denominated, which makes sense.

Sam thinks there is huge demand for Stablecoins and what came out of 2018 is the demand for a dollar-denominated (less regulated) product that moves freely between countries that do not have the best relationship with the USA.  Philip would put it differently: they are regulated in a different sphere of jurisdiction from the one underlying asset originated from.  Dollars are regulated by the USA, but the usage of these is happening in China and Russia, which are outside that sphere of jurisdiction to some extent.

Do you look at Tether flows? Yes, we do. One of the most interesting things for me in 2018, was when these Chinese mainland exchanges were shut down, a lot of this bitcoin was exchanged for Tether, and then this Tether moved off the mainland. This is why Tether is quite sticky as there was a lot of wealth transferred into during this period, and that formed the basis for that ecosystem.

Sam goes back and forth on Tether. You have to ask on the day. It does not have the best foundation and exists in an arbitrage structure because it exists outside of normal regulatory properties.  This is what creates the demand for Tether and why it is in demand in emerging markets.

Looking at Tether as an economist, what do you see? Philip sees essentially a gap in the market being filled. The reason Tether has been so popular is there was latent demand on two major fronts: (1) to have a stable pair in a trade (you don’t want to always be trading pairs that are correlated to each other), so there was a need for that, but (2) there was a restriction on Fiat getting into exchanges, and there was also a benefit for exchanges to remain Crypto to Crypto. Tether definitely filled that need for a stable fiat on-ramp and fulfilled that leg of the trade, and also there was latent demand for a dollar-like instrument in jurisdictions where people did not want to use the dollar all the time. That is where we see the demand for Tether coming from.  What I do think the fascinating thing about Tether is in its issuance and its monetary policy. I find it amazing that people are happy to abstract away from that and just say this is really useful. For me, this is a sign that there is the potential for even greater growth in Stablecoins because if people are willing to value it on par or sometimes worth more than a USD with the caveats that people have. Then there must be a whole bunch of people who would be willing to use a stablecoin that has slightly better properties. Sam completely agrees.

Tether proves that there is demand for a privately issued tokenised dollar that does not exist in an account in a bank somewhere.  I think this is what has driven Circle and other stablecoin operators and ramped up their operations.  There has been incredible growth in USDC in the past year.

In Sam’s head, the benefit of the company that is issuing the stablecoin, let’s use USDC, is that the underlying asset is a derivative of the cash balances that sit in their account. When you send money to Circle, they reinvest into bills or bonds so they can earn a yield on it. When you can be an asset gatherer, and you can earn 1-2% on that a year in relation to the stablecoin that you have issued, that’s a great business revenue model. But what happens if interest rates continue to go negative? It becomes harder for Circle to find the yield it needs to continue to issue stablecoins.  There are huge questions there.  Circle has come out with statements saying they will be okay, but that is just a statement.

It would be amazing if the best use case for Crypto was destroyed by expansionary monetary policy and negative interest rates. They would have to run their fiat investments with some risk to earn yield.  I don’t think you could treat it like money market funds because they would not make enough money to fund their business, especially with being regulated in the USA.  Let’s say they can go to 30-year bonds which are doing 0.5% at the minute. If the 30-year bond collapses any further, it destroys Circle’s business model. I am sure they will say they can continue, but it will see what happens in the future.  You can barely get yield for your dollars in the real world, but if you move into the crypto-sphere, you can get a much higher yield.

Do you think they could get to the point where you have to pay more to get a stablecoin from Circle, like $1.02 or above par further? Maybe.  The thing that is so interesting for us is that we can move between our bank account and stablecoins on a one to one par. The one thing I worry about with DAI is almost statistically there will be DeFi implosion because we are too early in the lifecycle to have all the kinks ironed out. We can definitely see the DeFi contracts are sending DAI to each other and are making loans on borrowed DAI. There is counterparty risk and leveraging going on in DeFi. Could we create those counterparty risk models? Who knows.

Are you doing analytical research into what is happening in DeFi at the moment? Not really. We work on a very large scale – Bitcoin, Ethereum, XRP – but in terms of commercial focus, it is not on our radar yet.

Is DeFi derivative leveraging?  There was a price collapse, and the volatility of Ethereum pushed the GAS price so hard that someone was able to take ETHER thanks to working contracts to take money from the CDP.

What were you looking at in Chainalysis a couple of weeks ago during that implosion? The top-level metric is on-chain inflows into exchanges.  The year to date figure up until 9th March 2020 was 52,000 bitcoin going into exchange daily. Then on 11th/12th March, you had 270,000/350,000 bitcoin respectively a day. Vastly higher multiples. Most of that came from people who had been holding bitcoin for 6 to 12 months. Those two figures tell a big story. There are many multiples of sell pressure coming in, and we know what Bitcoin markets are not incredibly liquid. Even if you get 10,000 more bitcoin coming in a day, you will see price effect.  That is why we saw such a huge price decline. We also saw an increase of smaller transfers on change, things less than a bitcoin. Retail is people moving less than one Bitcoin. We also saw a lot of large transactions too, which is more of the professional side of the market selling off.

Normally when you see bitcoin going into an exchange, the price effect happens 12 to 24 hours later because people don’t sell all at once.  They sold within hours of getting it onto the exchange. Things have been much quieter since these days in March, but there are still 100,000 Bitcoin a daily flowing into exchanges. But a lot of it is flowing off of Exchanges. There is another side taking it and putting it into storage. Then Tether pumped almost a billion dollars worth of new Tether into the system. We kind of have our own Federal Reserve who can buy up these assets. This seems to have stabilised prices.  Maybe there were professionals who needed to cover margin calls which created the selling pressure. The people who hold Bitcoin for longer periods of time; they didn’t move any Bitcoin. For now.

It doesn’t take very much to move the exchange, only about 5% can drop the price 50% given the metrics.  The spot markets are much more liquid than the derivative markets in Bitcoin.  The spreads were amazing to watch for a 12 hour period. It completely collapsed. The widening of spot versus Bitmex was something I had never seen before. There were quite large in-flows into derivatives markets on these days too, which was speculation as much as margin calls.  There were rolling margin calls all day which drives the flows.  Bitmex turned off the exchange and let the market reset. Crypto version of a circuit breaker I think.  Arthur just turned it all off at Bitmex. The derivatives exchange prices could have been pushed down to zero without the reset. I have some sympathy for Bitmex.

The on-chain metrics must have been really interesting. There is a big difference between being an Economist and a Trader – is my data right or wrong? How will it affect the market from a macro perspective? It taught me what people care about, and we were able to help clients make decisions during this event using data.   It was the first time that Chainalysis had enough data to actually make a description of what was happening using a data-driven narrative.

Would you say the fundamentals of Bitcoin are strong? I would say they are stronger than for any other crypto asset. I do pause because I do think there are differing visions of where Bitcoin should go, there is: should bitcoin be this financial asset that has widespread adoption, versus should Bitcoin be this asset that can be used by anyone around the world to do whatever they want? These are slightly competing narratives.  On the investment side, you need the mainstream to get comfortable with this, which means AML, KYC, regulated exchanges, regulation and custody. You need all these adult things. This does put pressure on the libertarian side of Crypto to be used for whatever you want. They can co-exist, but making this an investment-grade asset means more regulation and more compliance.

The reason for this digression is that we could be entering a new world after Coronavirus from an economic side. There will be a big recession, but after that, we may be exploring a new paradigm in terms of monetary and fiscal policy in the fiat world. It is unclear whether people will go, in a world of unlimited quantitative easing, we will see the effect of Coronavirus on economies, perhaps people will not want to look at something unfamiliar because of all the uncertainty and negative yields. I think we don’t know which world we are going to be in. Maybe that will create more space for using Bitcoin for whatever people what to use it for.

If you spend too much time on Crypto Twitter, you get narratives on the future. Your opinion is nuanced, which is much better than most people. Sam is a trader, so he looks at short term prices.  The biggest models you are seeing for Bitcoin right now are stock to flow. This just does not make sense; it shouldn’t be that each to 10x your money.  For Philip, the fundamental valuation model we have for Bitcoin right now are just not good. The main use cases for Crypto – holding and criminal activity – are not good.  These are not enough of use case for me at the minute.  I think that long term price forecasts need use cases that justify valuations.

The crypto ecosystem has not covered itself in glory during the last price drop. It is an illiquid market. We will see what happens, but it is a new paradigm when the 10-year yield goes to zero. At that point, the 10-year note becomes cash. What does that do to investment structures? There is this interesting point that when you take the volatility out of the market, you just displace it somewhere else.  At the same time, this volatility just gets pushed down the road, and here we are now. We can’t get away from the median before you need leverage to be reset. It is good for me, and for you, as young people but bad for our parents or those nearing retirement.

I was doing my undergrad in 2008 and went back and read a lot of the stuff I read then. I also thought a lot about the history of the last ten years.  It was not that fun graduating into that recession, and now here we are back into it.  I think the thing that I worry about is that macro….the monetary and fiscal policy we have had over the last ten years was actually inching us towards expansionary policies; that was because people were changing their minds about doing economic policy. Now we are sort of acting with all those lessons learned – they are dropping interests rates really fast and announcing QE really fast; however, there was no discussion about what this does in the future.

The reason I bring it up is because I worry that we hadn’t gotten far enough into the ‘future’ as it were, to learn some of the consequences of the policy decisions that were made over the last ten years. BUT we just double down on them; so I guess we will find out. I agree this is not good if you are about ready to retire, but I also think it will come back to haunt us when we are getting older. Potentially. If your portfolio is a mixture of bonds, stocks and equities (young people should have a 90/10 portfolio), that kind of portfolio could be really hurt by these expansionary monetary policies of the last ten years and going forward.  You need a diversified portfolio which includes alternative investments.  I understand why millennial investors would not want gold at the beginning. But I think this goes into the Bitcoin thing. Until recently, Bitcoin was an uncorrelated asset, and it benefitted your portfolio.  Over time, it does protect you from the perils of inflation.

From my perspective, maybe there were always large tail risks; but maybe we are just becoming aware of them.  Actually, before I got into Chainalysis, I did a lot on the economics of climate change. Climate change is all about how tail risks get fatter than people expect them to be, and that adds in more extreme events; people’s models do not take into account these events or their happenings. I spent a lot of time talking to large companies and government departments about how if it happens, you do need a large number of assets that function in extremes as well. We are in a very similar situation now.

What happens if Bonds and Equities, both of these assets decline at the same time? That kind of market is one we have not experienced yet. Maybe this is the situation, a huge decline in global demand leads to deflation in theory and should raise interest rates.  I think we are going to experience a huge increase in inflation and then you increase interest rates to reduce the amount of money floating around the system.  The velocity of money needs to slow down to bring inflation under control.

What happens when you have a levelling off of global population? All the models are built on the growth of population and as a result, GDP growth.  There is a point where our population levels off, at around 10 billion people. At that point, there is no global growth year on year.  Any inflation would create price rises but not any increase in demand.

How do you take into account these global growth economics into monetary policy and interest rates? I think the honest answer is that most economics don’t. It is so far in the future, and you discount those states of the world. I am not saying it is the right way to do it. You have discount rates because you want things now, not in the future; and there is essentially a ratio between those two things. Something that should guide your discount rate is the growth of the economy. As the economy grows, if you’ve got some stake in it, you should get some productivity growth out of it.  IF you are thinking of worlds with radically different growth rates, you can’t just add them up in one mathematical model – they need to be individually calculated these valuations based on radically different outcomes.  If you went back 15 years ago, that wasn’t thought about at all in Economics. The conceptual tools to understand this have been slow to develop. What people often think about in all long term economic models is that population is a given; there is a technological singularity that is baked into most modelling. I think the thing is that there may be scenarios where this might not happen!  People just don’t think about this. The assumption is that monetary policy doesn’t matter because the party hasn’t every stopped before. Coronavirus changes the party because it stopped the economy for three months.

It is the most damaging thing that the economy shuts down, but the debt just keeps piling up. There is nothing you can do other than forcing banks to give people debt holidays.  The next three to six months are going to be extremely damaging.  You look at the global leverage that existed before the first price drop.  It was massive. This is the first wave of deleveraging happening. In Sam’s head, this keeps going because deleveraging will continue. We haven’t seen the rest of it yet.

If Oil goes under $20 or under $10 for an extended period of time, there is so much debt in the oil industry that needs to be serviced.  Does this set back the development of renewable and alternative green energy sources?  In the power sector, power generation has basically been decarbonised. The power market operates on its own logic.  The natural gas price is also low. There is transport which uses most of the Oil, and that is definitely a challenge. What it is about is the stock turnover; what kind of car do people buy when they come to change over their vehicle? Fuel prices are important, but there are other factors that can be played with to change people’s preferences. It is not the end of the world; in 2013, it would have really changed things. There is enough momentum behind this decarbonisation to keep going.

Bitcoin mining, as a whole, uses huge amounts of energy to mine Bitcoin.  The carbon footprint of Bitcoin is the same as the entire country of Denmark.  It has 8.3-kilo tons of e-waste generation – which is the same as the waste generated by Luxembourg. It is not good for the environment. At the same time, humans do a lot of stupid stuff with electricity.  The environment impact of cryptocurrency is the Achilles heel of the industry.

Do other economists struggle to see why you want to work in the industry? They see why you want to work in Crypto academically, but those that work in complex systems totally get why it is interesting to work in Crypto with its data.  I do think it is pushing forward some of the research thinking in the fields of complex systems, but standard macroeconomists think I could be doing something better with my time.

How did you get into Crypto?  I had always followed technology, and I heard about Silk Road. I thought I need to learn about this, and this is going to be a thing. It changed a paradigm. It got me thinking about it.  John Levin, who was an intern at my previous company, founded Chainalysis, and he invited me to come and think about the datasets at Chainalysis.