The Fat Startup: the terrible incentives of ICOs


ICOs keep growing

I remember ICOs being a really hot topic last year. They were hailed as “revolutionary” and will possibly displace venture capital, but I haven’t heard as much about them lately. Someone recently told me that ICO activity has died down, so I decided to check to see if that was true. It turns out to be extremely false. In fact, so far in 2018 the average ICO is 50% bigger than in 2017 (though that might be skewed by the massive Telegram ICO), and halfway through the year they have already raised 140% of 2017’s total:

ICO Tracker

My impression for why the chatter around ICOs seems to have died down even as ICOs have increased is that companies are more cautious now about how they talk about it and are more likely to stay under the radar due to government regulators taking notice. Unlike the early days when ICOs were advertising aggressively (ICO advertising has now been banned on Facebook and restricted on Google), spamming inboxes, and making outrageous claims, companies today take a somewhat more sober approach.

The lessening of outrageous PR, however, has not made the underlying incentives for ICOs any better than they were before. No matter how you look at it, raising millions (or tens of millions or hundreds of millions) on no product and no users is rarely (ever?) economically justifiable and leads to extremely bad incentives.

In recent years, the “Lean Startup” method of experimentation, customer feedback, and prudent stewardship of money had become the standard in the startup world. This made sense – business plans are made up of hypotheses, not truths, and you have to test your hypotheses by building, releasing, getting feedback, and iterating in order to find product/market fit. ICOs are rolling back all the gains from Lean Startup insights and getting back to raising millions (or hundreds of millions) on business plans (white papers) alone. And fat startups lead to terrible incentives.

Why stay when you can leave?

Token whitepapers love to mention how they use game theory and economic incentives to revolutionize XYZ industry, but what they fail to mention is that if they indeed raised $XXX millions on no product and no users, they have no incentive to spend that money doing the hard work of making their whitepaper predictions come true – instead, the bigger incentive is simply to walk away.

I’m not even talking about straight up scams. I’m talking about teams that go into an ICO fully intending to realize their blockchain vision – they are still incentivized to walk away the moment things get difficult. Building a company is tremendously hard – getting rejections, managing conflicts, shattering your assumptions – none of that is fun or easy. Anyone in startups knows that trying isn’t enough – there’s a lot of grinding through, a lot of staring collapse in the face and being forced to come up with solutions that you didn’t think were possible a short while before. Part of what gets teams through the grind is the big payoff at the end – and a legal obligation to stay with it or return investor money. An ICO team that has raised $XXX millions on just a white paper doesn’t have those incentivizing factors because they’ve already gotten their big payoff, and they are not obligated to stay until all the capital is spent – they can walk away with it.

An ICO team may have good intentions. They may even try to tackle the problems that inevitably come up. And when they finally walk away from a project they may really believe that they did all they could, but would that be true? What if they could’ve succeed if they grinded for another few months? Or a few years? But what is the incentive to stay with the grind when you can walk away with millions of dollars and no real consequences? Founders may say at this point that this doesn’t apply to them because they’re “not in it for the money”. That’s bullocks. Money is never the sole reason for doing a startup, but it’s a pretty big incentive. To say that “money isn’t an incentive” is basically to undercut the whole premise of tokens, which is to use money to incentivize good behavior. Except in the case of ICOs, money incentivizes bad behavior.

Discipline, please

Let’s say a team decides to stay with the product until the money they raised runs out (or perhaps laws are passed compelling them to do so). Does that solve our problems then? Unfortunately, even then, their likelihood of success is probably lower than a similarly situated startup that raised money the traditional fashion. Outside the world of blockchain, discipline happens because professional investors and the demands of the market keep you accountable. At much lower raises, the need to make revenue and get to product market fit before your funding runs out acts as a disciplining factor that forces companies to focus on building the features most likely to create value. If you raise millions upfront, you can spend years in a basement somewhere building features without any care of whether it creates economic value for society or whether it justifies the capital you raised. That’s fine if you’re a scientist or researcher and your ICO specifies that you’re looking for donations, but most ICOs are advertised as investments that give a return.

People often think that startups fail because they run out of money. This is technically true since by definition a startup has not failed unless it has closed up shop (barring acquisition), and they usually don’t close up shop unless they run out of money. But what led it to that point in the first place? For a lot of startups, it’s because they raised too much money. The frank post mortem of this fintech startup hits on all of the problems I mentioned and more: when you raise too much money, you’re not forced to be close to the customer so you build things the market doesn’t want, you treat your company like a “science project” instead of a real product, having more money simply leads to spending more money in ways that are not value creating, you make suboptimal decisions that are not caught because money hides these issues, etc etc. The startup world is littered with such examples.

Some might argue that blockchain companies raise money not to build profitable companies but to build new technologies. If that were the case, they need to make clear to investors that the money raised is being donated to a research charity and that investors should not expect a return. New technology, no matter how innovative, is not automatically useful. The startup world is littered with even more examples of startups that failed because they focused too much on the technology at the expense of customers, and ended up with great technology that nobody wanted.


Some justify the need to raise large ICO rounds because 1) blockchain developers are rarer and more expensive so development on blockchain requires more capital than normal, and 2) there aren’t follow-on rounds in blockchain funding because there’s an ethos against dilution and token inflation. When I was starting out, I didn’t have a lot of money, so I asked a mentor: “should I hire someone cheap who isn’t very good, or should I hire someone expensive who is very good?” His reply was: “You should hire someone cheap who is very good, and if you cannot solve that problem, your startup will fail and you mind as well quit now.” The fact that costs are high doesn’t justify inflated valuations or change any of the bad incentives. And as for follow on funding, there is nothing stopping blockchain companies from raising future rounds. If a team shows real traction and their investors are actually committed to their success (which is supposed to be the whole point of tokens), those investors should be perfectly ok with companies raising future rounds as needed.

Here’s my prediction with ICO companies: they will fail at higher rates than traditional startups, even controlling for the maturity of technology and market. They will fail sooner because founders don’t want to put in the work when they can walk away with millions. And founders will stick with them for less time even if the startup hasn’t failed, because why stay and build long term when you already have your payoff – just leave the project to someone else while you chase newer and shinier opportunities. We already see this starting to happen.

Despite all the above, I’m not anti-ICO (surprise!). ICOs open up a new way of funding projects. This is great for a lot of projects that can’t get traditional funding (e.g. because the problem they’re solving is too big or not easily defined), for organizations that are indeed looking for donations for research and innovation, and for teams in countries with less developed VC networks. Just as crowdfunding opened up opportunities for new entrepreneurs and new innovation, ICOs will do so as well. But at least with the current set up, many if not most (all?) of the companies raising money this way are shooting themselves in the foot, and burning up investor money in the process. Speculate at your own risk.

What’s the point of Asset Backed Tokens?

ABT image

I went to an event a few weeks ago where people talked about asset backed tokens. I was curious as to what the point of asset backed tokens were if asset backed securities already existed. One speaker had a token backed by oil ( and another was looking into real estate backed tokens (

The speakers said that asset backed tokens allowed for greater liquidity of traditionally non-liquid assets like real estate, and allowed international and retail (small) investors to invest in assets previously unavailable to them. For example, instead of buying into a fund that requires a $1mil minimum or being shut out because of rules against international investments, now retail investors and international investors can also get in on the action, allowing for more funding for projects, more liquidity for investors, and more options for small investors in particular.

Funding and liquidity and options for small investors are all great things. But why tokens? You can already invest in illiquid assets like real estate and oil using asset backed securities like oil futures or real estate ETFs. To the extent that small investors are barred from certain investments it’s either because of regulation (such as those limiting certain investments to high networth individuals to protect small investors from losing money they don’t have), or because the managers of a particular investment find it operationally inefficient to keep track of many small investors. But the financial instruments and technology to allow for such investments have existed for ages. You don’t need blockchain to sell shares in your real estate fund or oil ETF. Companies like Stash allow users to invest in a wide range of opportunities with as little as $5, proving that it is both operationally and technologically possible to get small investors into diverse sectors without any use of blockchain. Timeshares that allow you own a small piece of a property have existed for a long time, with the added benefit that you can actually use the space (REIDAO, a real estate backed token, seems to be doing the same thing).

So what’s the point of asset backed tokens on the blockchain? It seems like there are two main reasons:

  1. Regulatory arbitrage: just a fancy way of saying “taking advantage of legal loopholes”. Right now, there is a dearth of regulation surrounding cryptocurrency, and blockchain companies are basically writing the law as they go. Oilcoin’s whitepaper mentions that Oilcoin plans to be treated differently under the law from other oil based asset backed securities, which gives tax and regulatory advantages. The decentralized nature of blockchain may allow skirting national laws limiting real estate investment to citizens. And uncertainty around what exactly a token is may allow companies to evade regulations on securitizing assets or skirt laws that limit investments to only wealthy individuals.
  2. Stablecoin: right now there is a lot of volatility in the cryptocurrency market which prevents widespread adoption of cryptocurrencies as real money usable in every day transactions. There’s been a lot of talk about the need to create a stable coin that isn’t massively fluctuating each day so it can be used as a practical medium of exchange and a stable source of value. Backing a token with stable assets can lead to stability in the token and allow it to be used as a stablecoin.

Regulatory arbitrage

“Regulatory arbitrage” is the practice of finding loopholes in the law to achieve some gain. Wealthy people using offshore tax shelters or companies keeping overseas profits abroad so they don’t pay taxes on them are examples of regulatory arbitrage. Profiting off of regulatory loopholes is probably as old as regulation itself and is a significant factor in the growth of cryptocurrencies. Most ICOs are “allowed” to raise millions of dollars with almost no disclosure or investor protections only because laws in this area are unclear. Many US based cryptocurrency companies are registered overseas to provide an excuse to not follow US disclosure requirements and participation limitations, making it much easier for these companies to raise funds. At any point, however, US regulators can tighten these loopholes, e.g. by saying that ICOs are like IPOs and therefore companies that ICO need to follow IPO requirements. Regulators can go further by cracking down after the fact on companies that have already ICO’ed, saying that they should’ve known to follow IPO rules, and fining or even shutting them down. This is why the more legally cautious major ICOs like EOS are not available to US buyers, or are only available to accredited (i.e. wealthy) investors. Many ICO companies and investors, however, are working regulatory loopholes as hard as they can while the law is still unclear.

The asset backed tokens I’ve encountered are not as flagrant in arbitraging regulation as purely digital tokens. Because asset backed tokens involve real world assets and not just “imaginary” cryptoassets, they have to at the very least, comply with existing laws on how real assets must be priced, stored, transacted, etc. Real estate tokens, for example, deal with complicated laws around title ownership, leading to some, like Praetorian Group, to wait for SEC approval before moving forward with their ICO. Oilcoin takes care to make clear that that they will comply with laws governing commodity transactions, financial disclosures, and “applicable tax laws”. It simply interprets those tax laws in a novel fashion that gives it an advantage over traditional oil backed securities.

There are two ways to look at regulatory arbitrage in the context of blockchain – one is that this, like traditional forms of regulatory arbitrage like tax shelters, is a form of “rent seeking” – a way of making money not by creating real economic value but by shifting the way value is distributed. It’s taking a bigger slice of the pie rather than making the pie bigger for society. The second way of looking at this is that certain regulations themselves are outdated and prevent value creation in a rapidly innovating economy. By pushing the boundaries of regulation, cryptocurrencies force regulation to adapt and this is in fact value creating for society as a whole. For example, creating a Real Estate Investment Trust, the current way to invest in real estate backed securities, is a complicated and expensive process, whereas creating real estate tokens may be only a fraction of the cost (especially if you avoid many of the regulations). Only time will tell which explanation is the truth. If 10 years down the line we find that ICOs have, on the whole, created wealth for small investors, then we know that the boundary pushing today is value creating and that current regulation is outdated; on the other hand, if 10 years down the line most cryptocurrencies have failed and most investors lost money, then we know that regulation today should have been more restrictive, and that boundary pushing ICOs were not value creating but simply shifted value from naive investors to blockchain “entrepreneurs”.

I think that while most cryptocurrencies engaging in regulatory arbitrage do so for rent seeking reasons, the reason that regulators have not cracked down harder is partly because they recognize the potential need to update regulations for a new era and want to see how the token market evolves before deciding whether and how to pass new regulations. I think that is wise. The blockchain case is different from old fashioned tax evasion because this new technology has the potential to create innovation in business models and economic organization, which would in turn require new laws and regulatory schemes. So even if individual token companies are rent seeking,  pushing for new tax treatment and new investor regulations as a whole may very well be value creating as they force regulators to look harder at this new technology and think about how to regulate it in an innovation enhancing way.


A possibly more interesting use case for asset backed tokens is as stablecoins. Many people see cryptocurrencies as a secure and trustless (as in you do not need to trust banks or governments to validate your money) replacement for fiat currencies. For that to happen, tokens need to be stable enough to purchase real world goods rather than fluctuate wildly due to speculation. Most tokens do not have any “fundamentals” behind them – unlike equities, their value is not based on real economic factors like income and operations, and unlike fiat currencies they’re not backed by national economic trends. Since most cryptocurrency projects thus far are pre-product and pre-revenue, the value of their tokens is purely speculative. Backing tokens up with real world assets grounds its value in economic fundamentals, which should make the coins more stable and more usable for real world transactions.

Many asset backed tokens plan to step into that role as stablecoin. Praetorian Group is planning to issue a debit card that uses its tokens for real transactions, with the assumption that its tokens will have stable value based on its real estate assets. Digix seeks to create a stable coin backed by gold. Tether is backed by fiat currency (US dollars), there are even some backed by other crypto tokens. Venezuela is also planning on a token backed by its oil reserves to replace(?) a national currency broken by inflation.

New stable coins seemingly pop up every day. Some have called it the “Holy Grail” of cryptocurrencies. An investor joked to me that he is open to every blockchain project, unless you were “the 12th stablecoin to talk to me that day”. I personally think that a true stable coin should not require backing by real world assets and should be pegged to the price of some basket of useful goods (e.g. food and household necessities). This is what companies like Basis are trying to do.

Criticisms of ABT

Besides the criticism of ABTs as “rent seeking” mentioned earlier, the chief criticism of asset backed securities is that they are not decentralized. Gold and oil must be stored in physical locations. Real estate exists in a few discrete places. Fiat backed tokens (e.g. those backed by USD or Euros or Yen) must store their dollar reserves in real world banks. This seems to defeat the purpose of having decentralized cryptocurrencies in the first place, which is (in significant part) security and trustlessness. What makes cryptocurrencies secure is that they simultaneously exist everywhere, so there’s no central place to attack and disrupt the currency. But if you store all the assets backing that currency in a few centralized locations, then someone can simply attack your storage vaults to disrupt the currency. Decentralization also means you don’t need trusted intermediaries (like banks) to validate balances and transactions, because the whole network stores your balances and verifies your transactions. But when tokens are centrally stored, this means that you need to trust the organization storing them, which makes it fundamentally the same as trusting banks and other centralized institutions.

To evaluate this criticism, it’s important to recognize that there is a foundational difference between asset backed tokens looking to be stablecoins, and asset backed tokens that simply use blockchain as a way to invest in the underlying asset. One represents currency while the other represents assets. You use currency as a way to transact goods, while you invest in assets hoping that the value of those assets will increase. It’s like the difference between a dollar and a share of stock.

The point of a stable coin is to be a currency that provides all the security and trustlessness benefits of the blockchain while still being a stable medium of exchange for real world goods. So actually having the positive properties of a decentralized blockchain is important, otherwise what’s the point of using the token over fiat? For this reason, I think that a true stable coin should not be backed by assets that require central storage and trust in the storage organization. This doesn’t matter as much for asset backed tokens that exist as investment vehicles (for oil, real estate, etc). These tokens are fundamentally asset backed securities, and they exist to provide increased investment options at lower cost with more liquidity, not to provide some sort of secure and trustless fiat replacement. The focus of these tokens is the underlying asset you’re investing in, and blockchain is simply a tool to better invest by reducing operational costs and arbitraging regulation. While some tokens, like the aforementioned Praetorian Group, is trying to play both the stablecoin and investment vehicle roles, these roles have fundamentally different purposes and requirements.


For those thinking about investing in asset backed tokens, make sure you understand what their underlying value-add is. Are they looking to be a stable currency exchanging between the crypto- and real world economies? Or are they an investment vehicle for real world assets? Investment-vehicle tokens don’t really need to be trustless or decentralized. You evaluate them as you would any traditional investment – whether you trust the fund/token management group, whether you believe in the underlying asset, whether the return vs fee ratio is worth it.  If you’re looking for a token to use as a stable hold of value or medium of exchange, its ability to actualize the unique properties of blockchain like decentralization matter much more.

Practically speaking though, even for stable coins, you don’t need them to be perfect to gain some of a blockchain’s benefits. For example, Tether is a USD backed token used to transact fiat currency across the blockchain. The Tether token isn’t intended to buy goods but rather to be used as a temporary waypoint in fiat transactions. This allows you to gain the speed, immutability, and low cost benefits of sending money on the blockchain, even if you don’t (fully) gain the decentralized and trustless benefits of the blockchain.

As blockchain and cryptocurrencies evolve, it’s likely that various tokens and services will realize some of the benefits of blockchain without realizing others, so the utopian vision of a perfect coin for a perfect blockchain economy may not be necessary for  society to enjoy the varied advantages that blockchain brings.


Do tokens incentivize good governance? A look at AdChain

Strong assumptions

One of the themes I come across when reading about blockchain is the idea that tokens  incentivize good governance and counteract bad behavior: because tokens are valuable only if the product/platform that issues it is valuable, token holders will actively participate in the governing and development of the issuing product/platform in order to keep the value of their tokens high. My question is, how robust is that assumption, especially when we consider free-riding and the concentrated costs/dispersed benefits problems that we routinely see in the regular economy?

The specific case that I’m going to consider is an interesting blockchain company called AdChain. AdChain was built by MetaX in conjunction with blockchain behemoth ConsenSys and industry group Data & Marketing Association. They look to use token economics to tackle the multi-billion dollar problem of digital advertising fraud, where websites use bots to create fake “impressions” to scam ad dollars from advertisers. I first encountered AdChain when one of their team members spoke at a blockchain summit I attended in March 2018, and I was very intrigued by the problem they were solving as well as the economics embedded in their model. They had an ICO in June of 2017 and raised $10 million in a day, and went live on the Ethereum Mainnet in late April 2018. I’m not going to evaluate their tech, their team, or their business model, other than to say that they are indeed tackling a very important problem. Instead, because they are such a perfect example of a token economy making many of the common assumptions that other token economies make, I want to unpack the economics outlined in their whitepaper and think about the implications.

What’s a rational token holder to do?

Basically, AdChain is a whitelist of trustworthy websites for advertisers to spend ad dollars on. Websites apply to be listed on Adchain by buying and then staking adTokens. Existing token holders can challenge an application by staking their own tokens, which then initiates a platform wide voting period where other token holders vote on whether to admit the applicant (votes are token weighted so the more tokens you have the more votes you have). The winner (the applicant or the challenger) gets the tokens that the loser had staked, with some proportion going to the winning voters as well.

They key assumption here is “applications for apparently fraudulent or low quality domains will be challenged by rational adToken holders” (adToken white paper, p. 3). This assumption is based on the idea that adTokens are valuable only if AdChain is a credible whitelist that advertisers want to use and websites want to apply to, so to keep the value of adTokens high, token holders will be diligent in weeding out bad applicants. This is  a specific case of a general assumption in the cryptocurrency community that holders of tokens will be incentivized to actively govern and participate in the community. The question for us is, how likely is this?

Concentrated costs, dispersed benefits

The “real world” example most analogous here is probably stocks, where holders of stocks also get to vote on certain aspects of the company, with the expectation that stockholders will participate in the governance of a company to increase the value of their stock. In corporate governance however, it’s long been known that for many shareholders it’s actually more rational NOT to actively govern but to free-ride off the efforts of others, because governing is expensive to you, but the benefits of your governance is dispersed among the whole community. This problem can be partially mitigated by having large shareholders, but even so, discipline is rare, and the presence of many short term speculators who have no interest in long term governance will further weaken oversight. AdChain tries to mitigate this by offering a significant reward for active governance (the winner gets the loser’s staked coins) – this is a great idea, and it’ll be interesting to see whether it’s enough to overcome free-riding, since even with this “governance bounty”, much of the benefits are still dispersed.

Another potential challenge facing this token economy is the asymmetric incentives between applicant websites and token holders. There’s a well studied phenomenon in political science where small special interests are able to defeat policies that benefit the much larger society because the costs of the policy falls on a concentrated group while the benefits are widely dispersed. US sugar subsidies are a prime example, where ending the subsidies will save consumers $2.4 billion per year, which amounts to less than $10 per individual consumer but will be a big loss for the sugar industry, so the sugar lobby is much more motivated to keep the subsidies in place than consumers are to eliminate those subsidies.

In AdChain, the incentives of fraudulent websites to cheat the system and get on the whitelist may be higher than the governance incentives of token holders. The benefit to the platform of preventing one bad website from sneaking in is dispersed among all token holders, but the cost to that rejected website is concentrated. What’s at stake for a website applicant is not only the tokens it stakes but its future ad revenue, so it has more at stake than token holders who might challenge them, giving website applicants more incentive than potential challengers to spend resources mobilizing voting blocs. That said, even fraudulent websites would have an incentive to maintain some level of integrity to the whitelist, so it would be very interesting to see how participants actually behave and what kind of balance, if any, can be reached.

Not a problem?

Practically, the problems of dispersed benefits may be mitigated in AdChain’s case and in the case of many blockchain startups. A “large” shareholder in a public company might have as little as 10% ownership or even less, with thousands of other shareholders holding tiny pieces of the company. Contrary to the idea of widespread democratic participation that many people have of ICOs, AdChain’s ICO, like many ICOs, had a small number of participants (“whales”) who bought large stakes, with nearly 40% of tokens publicly offered going to a single buyer. This means the benefit to keeping the platform safe is in fact quite concentrated. That said, we do not know how dispersed ownership will get once tokens are floated on exchanges. Depending on how many of these whales are speculators looking to dump vs real investors in the platform, the effectiveness of AdChain’s economic model may vary greatly.

AdChain went live only recently. As with most startups, it’s hard to say how the product will turn out simply by theorizing ex ante. The token economy may in fact be a completely new case with no comparable analogies to the current economy. And with all good startups, models and assumptions can change on the fly. Six months after going live, every blockchain startup will probably end up drastically different from their white papers, just as regular startups differ greatly from business plans once they launch. But to the extent that blockchain companies continue to raise funding via token offerings before they have visible traction, and given how important token incentives and governance is, it could be helpful for token investors and entrepreneurs to consider the possible limitations of token economic models the same way they consider limitations in technology models.

I, for one am rooting for this startup and other blockchain startups tackling important real world problems. I’m excited to see how these mini economies unfold as more companies go live and put their assumptions to the test.


Making the economics of cryptoeconomics a little less hidden

About Me

I hold an AB in Economics from Princeton University and a JD and MBA from Columbia University. I started my career in economic consulting, and I currently run a boutique digital innovation studio based in NYC, where we consult for clients and do research in new technologies:

Purpose of this blog

I got interested in blockchain in 2017 when I found out about its combination of economics, law, and technology. I thought, finally! Law and economics can be directly applied in an innovative technology context! I started reading up on the foundational technology of blockchain, and also devoured white papers on new blockchain and cryptocurrency startups, mainly those with a strong economic (rather than just technical) component. As I read, however, I found that while many blockchain startups were based around token economies and creating new economic systems, the portion of their white papers devoted to how their token economy and incentive structures work were often tiny compared to the portion devoted to their technology. In fact, in many papers the economic assumptions are not even spelled out, hidden among conclusory statements and technical jargon.

I think a more thorough analysis of the economics underlying various parts of the blockchain ecosystem – token economies, ICOs, mining incentives, etc could be helpful for the entrepreneurs involved to think through their business, for ICO investors looking for solid projects, and for technical talent considering joining blockchain startups. This blog will analyze the economics of blockchain and interesting blockchain/cryptocurrency startups that I come across, speaking both theoretically and using case studies to hopefully help unmask some of the “hidden” parts of cryptoeconomics. This blog is not meant to criticize though it may be critical, and I certainly don’t pretend to be an economic oracle so don’t blame me if you pass on an ICO that then goes up 1,000x 🙂

I look forward to exploring this space together!

Good company in a journey makes the way seem shorter. — Izaak Walton