#190 Hubris Leads to Downfall

Global Policy Watch #1: FinTech Manoeuvres

Insights on global issues relevant to India

— RSJ

One of our favourite topics to talk about around here is regulations. We aren’t dogmatic about things. But the one principle that comes close to being a dogma for us is our belief in spontaneous order. The world is a complex interplay of many economic and social networks. Don’t try to force an order on it. Let the spontaneous, uncoordinated actions of the millions run their course. Some order will emerge from it. There might be occasional ‘disorder’ but inbuilt into such systems is an autocorrecting mechanism which will kick in. This is better than some powerful entity (like the State) trying to force order because it thinks it knows best. No one really can be a Laplace’s demon. The top-down forcing of an order will make things worse. 

That said, we aren’t libertarians by any stretch. The State must intervene when there’s a market failure. And more than any other sector, I believe financial services need regulation because of two obvious market failures endemic to them.

One, there’s a serious information asymmetry between the supplier of these products and services, and their customers. The products get more complex over time as suppliers look for additional arbitrage to make higher income, and customers can only understand so much of what’s often non-intuitive. I mean, compounding as a concept is a leap for most people; figuring out more complex instruments ain’t going to be easy. Two, the market power and dominance gets built up really fast in this sector. With it comes the risk of a contagion or the spectre of ‘too big to fail’. All of these leads to misallocating of capital that’s worse than most state interventions. So, I tend to look at financial regulations somewhat more benignly than, say, regulating cattle transportation. 

Having got that preamble out of the way, here’s a news item that many of you might have seen last week:

“The founder of the world’s second-largest cryptocurrency exchange, FTX, has apologised for his company’s near-collapse this week, saying he “fucked up” in his calculations and in his communications during the crisis.

Due to “poor internal labelling of bank-related accounts”, he said he “was substantially off” in his calculations of the sums that the exchange had lent out to users to let them make leveraged bets – borrowing money to trade with, magnifying potential gains and losses…

The sudden collapse in value was prompted by leaked documents which implied that Alameda Research, a hedge fund tightly intertwined with FTX through its common owner, Bankman-Fried, was, in effect, insolvent….

The leaks about Alameda turned into a crisis for FTX when Binance, the largest cryptocurrency exchange, announced it would sell its own major stake in FTX. The fire sale that followed crashed the token's value far below the $22 floor that FTX had committed to support and prompted the equivalent of a bank run at FTX itself as customers raced to withdraw their deposits faster than the exchange could process them.”

It wouldn’t be out of place to say there’s a lot to unpack there. But is there really? I have held a general theory about fintechs for a while. There are three sources of value in fintechs as I have seen them. 

One, they build slick, frictionless journeys that make buying financial products easier and more intuitive. Traditional players either don’t think this way or their legacy systems just don’t allow them to create such convenience. This focus on customer journey is good and somewhat sustainable in sectors where customers value speed and efficiency over risk management. Now ‘friction’ isn’t great in most cases but it has a few benefits. It slows things for you to pause and reconsider. In financial services, some friction is necessary to protect the customer. Not all of them have the ability to manage speed and the attendant risks that come with such lack of fiction. Anyway, frictionless works in sectors that have savvy customers. The best examples of such disruption have been new-age retail brokers like Robinhood, or closer home, Zerodha. There isn’t a lot more there in terms of product innovation or better management of risk that are traditional sources of value. It is all about convenience. Of course, such simplification and gamification - like, animated confetti floating all around when you make an options trade - runs the risk of drawing customers who don’t have the financial nous or the appetite for such risky products. And those accidents happen. But you could still argue there’s some tangible value fintechs have created in forcing everyone to rethink customer journeys in the industry. Is that sustainable, and does it warrant a huge valuation? Not quite. But at least you’ve got a real reason for customers coming to you. 

The second source of value - risk arbitrage - is of a more dubious kind. Traditional finance is built like it wants to reject customers. You must know your customers well, assess their risk profile, check for the suitability of the product and then offer them your product. At least, that’s what is expected. And for the risk appetite you decide on, the regulator will ask you to maintain capital so that you don’t ever have to turn your pockets inside out, tilt your head to the side and tell your customers you have no money. If you slack off on this, you might get away for a short while, but the chickens come home to roost fairly quickly. You have stringent disclosure norms, so there’s nowhere to hide when things go wrong. This means traditional finance does the equivalent of an airport strip search for every customer while taking them on. Sure it inconveniences 99.99999% of the customers but there’s no point taking that infinitesimally small risk by going around and trying to avoid it.

Now, this isn’t exactly how most fintech models operate. There’s greater value in getting truckloads of customers in really fast, offering them discounts because no one is asking you to make profits and creating a Bezos-like ‘flywheel’. This model militates against prudent risk management practices. The many lending apps that promise loans in seconds, the entire Buy Now Pay Later (BNPL) craze that swept through in the last few years and the many variants of such models are good examples of this mindless model. The only good news is that such models are up against fundamental economic sense. The risk starts to bite soon, and there’s nowhere to hide. The dismal performance of such fintechs once they’ve gone public should surprise no one. The only way out is by going back to the harsh reality of doing business in finance - manage your risk. 

And that brings us to the third, perhaps the rarest, source of value among fintechs - product innovation. This is rare because of two reasons. Firstly, traditional finance is innovative enough. This might seem surprising, but I guess a charge that’s laid on banks is that in the last thirty years they’ve innovated too much. It led to what’s often called the financialisation of the economy. In fact, one of the lessons from the global financial crisis (GFC) for the regulators was to clamp down on innovation to reduce the possibility of toxic products. So, there’s not been any dearth of innovative thinking in traditional finance.

Secondly, real product innovation is hard work. It requires deep knowledge of the domain and an understanding of the customer to arrive at one. Now, there aren’t more than a handful of fintech models that truly qualify as innovative, or to apply that much-abused term, disruptive. And if we have true product disruption, then there will be an information asymmetry-driven market failure deeper than usual. I mean if people don’t understand the run-of-the-mill financial products well now, what are the odds they will appreciate a real disruption? What this means is that ‘disruptive’ regulations will have to be applied to address the disruptive model in play. Importantly, the timing has to be right. Apply them too early, and you won’t have the disruption that’s useful in the long run. Go too easy on them, and you might create a beast that will be difficult to tame when things go wrong. 

Crypto(currency) is a real product disruption by this definition. We have covered it in a few previous editions (here, here, and here). The promise of a decentralised monetary system that isn’t ‘repressed’ by central banks worldwide has its own appeal. Not to me, but there are all kinds of people in this world.

And FTX was (as I write this I hear it has filed for bankruptcy) a fine example of the possibilities of this disruption. It was the second-largest crypto exchange in the world. You could open an account with it and transfer some money to it to buy any bitcoin. And they would hold it for you. When you want to liquidate the bitcoin for money, they will do so. They will charge you a small transaction fee for their troubles. That’s it. Not much of a business model till you start complicating things. Bitcoin isn’t exactly a stable asset. Its price fluctuates quite wildly. Now to some, this is a problem, but to people in financial services, this is an opportunity. So FTX started offering other services. You could borrow money from them to buy more bitcoins, or they could offer you bitcoins for money if that’s what you wanted (maybe you wanted to short bitcoins). Soon, it was no longer just an exchange or securities brokerage business. It was doing (margin) lending and taking risks on its books. It also started issuing tokens (think of it as its own kind of currency) that it offered to people to lend against. One of the things about finance is that there’s no real product that’s shipped from factories. People can make up products on the fly if they spot an opportunity to arbitrage time or risk appetites. So things get out of hand in terms of complexity soon. With FTX that’s what happened. You are lending and borrowing across time and risk horizons and you don’t know whether you can have your assets and liabilities match over the medium term in all likely scenarios. This can get messy in a traditional bank with all the bells and whistles of risk management and treasury teams monitoring this daily. In a fintech start-up that has some opaque tech platform with limited supervision and scenarios planned, it can be quite catastrophic.

In FTX’s case, they had the added complexity of having their own tokens, which they were accepting as collaterals against loans. So FTX had loaned a load of its customer assets to another of its company (Alameda) in return for its own tokens. The value of the tokens was based on FTX’s value. But FTX was also sitting on a load of them as collaterals.  If the value of the tokens fell, FTX would be impacted because it held a lot of them. If FTX got impacted, the tokens would fall further because that’s what the intrinsic value of the token was. If you are a token owner worried about the token crashing, you would take out your money asap. And if everyone thought like you, as they will, you will have a run on the bank (or the exchange). It was tailor-made for a death spiral. All it needed was a spark which was provided by its friendly rival Binance who tweeted one fine day about what it thought of FTX’s tokens. 

FTX, which was valued at $ 32 billion about a month back has filed for bankruptcy. Its investors (like PE giant Sequoia) have marked their investment value down to zero. Some customers will almost lose all their money here. There are counterparties involved, so we will know the full extent of its collapse after a while. The details that have emerged since about how FTX was governed are hilarious. There’s no other way to put it. There were only three members on its board. Two executives, including the founder and a lawyer based in a tax haven. No equity investor had a presence there. There was hardly an Asset-Liability Management (ALM) committee or a model. The founder’s partner ran the finance function. The whole thing ran on the back of an engineering team that built the platform with hardly any supervision on what was coded. The founder tweeted that there may have been a ‘labelling’ error in one of its fields and that could’ve led people to believe it had a liquidity problem. I used to make such excuses for MS Excel errors at the turn of the millennium when I was wet behind the ears analyst working for peanuts. For all the talk of software eating the world and decentralised platforms disrupting governments, it would be good to begin by acknowledging that the generations that have gone before us building what they built were no fools. In financial services, regulations are often a feature, not a bug. Thinking people don’t get it because they don’t understand disruption has no basis. It is hubris, and that leads to a fall. That’s the only lesson of history. 

India Policy Watch: Don’t Choke Cheap Chips from China

Insights on domestic policy issues

— Pranay Kotasthane

A lot has been said about India’s dependence on chip imports. Especially on chips coming from our long-term strategic adversary, China. And yet, I haven’t come across an analysis to understand the composition, size, and trends in this particular item of trade. So, I decided to investigate it further using the Department of Commerce’s Trade Statistics portal.

Chips (or Integrated Circuits) are India’s eighth-biggest import item by value. Energy imports — petroleum, petroleum gas, and coal rank above it. So do gold and diamond. The other items we import more than chips are products made using them — computers (including laptops, desktops, printers. etc.) and telephones (including mobiles, modems, etc.).

Now, let’s look at the chip imports from China. Considering PRC and Hong Kong as one political unit, India imports 64% of its ICs by value from China. Over the last five years, imports from PRC have increased seven-fold. Check this chart below to see the various kinds of ICs that come to India from China.

While this chart may look alarming, it can be misleading in many ways.

One, the absolute increase in chip imports from China follows the overall trend of increasing chip imports to India. Total chip imports to India have also increased sevenfold over the last five years. India is importing many more chips from all countries, including China.

And this is a good thing. It’s positive because it indicates that a lot more electronic device assembly is happening in India. A country will import chips — not finished electronic goods — in large quantities, only if it has a downstream domestic equipment manufacturing ecosystem. As even more Samsung, Apple, or Redmi devices get assembled in India, we should expect chip imports to India to rise for at least another decade in the best-case scenario, by which time we could have a semblance of chip manufacturing done here.

Two, chips imported from China are not necessarily Chinese chips. For one, it is likely that they are just assembled in China. Chip assembly is a labour-intensive process that is outsourced to other companies. China is a much bigger player in outsourced assembly and packaging of chips, than in fabrication. So, it’s quite likely that a die (an unpackaged chip) travels from Taiwan or Japan to China, gets packaged and finds its way into India. Secondly, even when chips are fabricated and packaged within China, a portion of it is done by foreign companies with facilities in China (such as Samsung, UMC, SK Hynix, etc.).

And three, its a fool’s errand to think we will ever be in a situation when no chips are imported. If India’s own chip assembly and packaging takes off, import of unpackaged chips will rise substantially. ICs are in fact the biggest import category by value even for Taiwan, the semiconductor superpower. It’s a folly to look at the entire semiconductor supply chain through a “national security first” lens. Cross-border movements of chips are desirable and inevitable.

To mitigate risks, what can be done is to diversify the source of India’s chip imports. And this is where the US export controls on China might have some effect. Though these export controls target only specific high-end chips, Chinese fabs will increasingly be pressed into meeting its domestic demands. Moreover, the confusion over further decoupling means that equipment makers will look at a China-free supply chain.

So, a significant portion of our chips are indeed imported from China. And yes, importing cheap chips from China is alright.

Applications for the re-awesomed Post-Graduate Programme in Public Policy are now open. Check details here.

Global Policy Watch #2: Seasonal Bird Migrations

Insights on global issues relevant to India

— Pranay Kotasthane

Exactly three years ago, I created an account on Mastodon, a decentralised social media platform that aimed to challenge Twitter. But the network effect pull of Twitter was just too strong though. I didn’t post a word on Mastodon, and as usual, forgot my login password.

But with the new Twitter boss taking sudden—and some inexplicable—decisions, another round of en-masse flight to Mastodon has begun. The 2019 en masse migration was India-centred, after Twitter blocked the account of a well-known Supreme Court lawyer. This time however, the migrants are from across the world. It’s the kind of natural experiment that social scientists turn into papers and books. I’m trying to be more active this time around, and observing how this new digital community takes shape. Having spent a few days, here’s what I have noticed.

One, the importance of the “other” doesn’t disappear. Most of the posts in my corner of Mastodon are self-referential. They are criticisms of Twitter and Elon Musk, and explanations for why Mastodon is better. Many early movers detest Big Tech, Capitalism, and such like; some are in love with “-isms” from the left-liberal political network. Beyond a point, the utility of Twitter as the “other” will decline, and I wonder what would it be replaced by. Finding the “other” is a core aspect of human behaviour. While the federated nature of the platform allows people to block people or communities (servers) easily to prevent confrontations between two ideologically distinct communities, I doubt it is powerful enough to overcome human desire for dissing an out-group to build in-group solidarity.

Two, if this is Web3, it feels a lot like Web1. This is a Web3 avataar minus the blockchain, bitcoin, and virtual reality. Not everyone across the world is in one town square. You need to know the coordinates of another person to be able to talk to them. In essence, it’s a lot like Web1. Many old-time internet citizens are celebrating this freedom from the powerful intermediary-controlled Web2. However, most users who were born into the Web2 might find this retro version inexplicable. For that generation, open social media platforms were never about “healthy discussions”, as the Web1 folks like to remind. Open social media was always something more instrumental—for a job, to access instant news, or to build an online persona. For discussions with friends, people used encrypted peer-to-peer platforms. So, it will be interesting to see how this generation of people finds a platform where non-virality is a feature and not a bug.

Three, there’s a large-numbers paradox at play. The initial few days on it have been quite good in this respect — the users seem to be warm, friendly, and welcoming. As long as it remains a small, elite group of people with similar worldviews, it might survive. But if it does get popular, and more people from various ideologies and with differing motivations join in, it is could again become just as vicious yet attractive medium like Twitter. While the architecture is aimed at preventing the slide into another big, loud, and heated forum, human predilections are the biggest bottleneck.

In any case, I’m going to observe this new experiment for some more weeks at the very least. I’m on Mastodon at [email protected]. Connect with me there if you are in another corner of the fediverse. Let’s see how far this experiment takes us.

HomeWork

Reading and listening recommendations on public policy matters

* [Book] The Aeroplane and the Making of Modern India by historian Aashique Ahmed Iqbal explores the link between technology and sovereignty.

* [Podcast] Michael Munger and Ross Roberts discuss a topic that’s on all governments’ mind - Industrial Policy - on EconTalk.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit publicpolicy.substack.com