Class summary review

What this is for

The slides are the primary source, and they are largely pictures: a map with two telegraph lines, a warehouse in New Jersey, a diagram of a stock trade with a dozen boxes, a girl facing a bull. Pictures work in a room with someone talking over them. A week later, on their own, what goes missing is the link: why that slide follows this one, why a crossbow is in a finance course, why a bicycle. This document supplies the links. Read it with the deck open, or on its own. For the long version, slide by slide with the classroom discussion, see the Lecture Companion.

On numbers. The figures on the slides give magnitude and direction: big or small, rising or falling, ahead of us or behind. Read them for that, and spend your attention on why they matter.

The argument in one paragraph

Finance is expensive, and it has stayed expensive. For roughly 130 years the cost of financial intermediation barely moved, about two cents a year to create and maintain a dollar of intermediated assets, while the technology underneath it was transformed several times over. Meanwhile the people inside the industry were paid steadily more: roughly level with the rest of the economy in 1980, and about 70% more by 2006. The research finds little evidence that better markets or better allocation of risk explain that premium, which leaves substantial room for rents, meaning pay for the position rather than the work. A high-cost industry earning high rents attracts entry, and technology has been removing exactly this kind of barrier for three thousand years. FinTech is the current instance of a very old process. The module tells that story three times.

The three waves

v1.0: technology has always reshaped trading

In 1850 the telegraph ran Paris to Brussels and Aachen to Berlin, with a gap between. Paul Julius Reuter flew pigeons across the gap and sold Paris prices in Berlin hours ahead of the train. The reason anyone would pay for that is arbitrage: information travelling faster is money, and where there is a rent to be made, someone will find a way to extract it. When the line was completed the pigeons were obsolete, Reuter moved to London, and the arbitrageur became the infrastructure. High-frequency trading is the same story 150 years later, and it disrupted the incumbents the same way, picking off the slow institutions until they were forced to modernise.

Trading automated first because the gain was largest and most immediate. A scarce human skill, drawn from a narrow demographic, was replaced by programming and statistics, and whole career streams disappeared. But markets got better: the post-earnings announcement drift, an anomaly taught in every finance course, has largely disappeared in recent data. That sets the tone for the whole course. Technology improves markets. In class the same point came with the Canadian version of the pigeon story: the TMX data centre is in Markham and a rival venue used to sit downtown, and that gap was arbitraged exactly as the Brussels-to-Aachen gap was.

v2.0: what changed after 2008

Two things, of different kinds. A crisis that destroyed savings and trust and created demand for an alternative, and a distribution channel: within a few years everyone had a computer in their pocket, and it was a supercomputer by any earlier standard. The catalyst was not financial. Everything else moved onto the phone first, and then the pandemic made it intolerable that banking had not: if the phone can get dinner delivered, why can it not do your banking? The banks took about a decade to react, and survived taking a decade.

Disruption comes from outside because the industry is self-contained: it hires a certain type of person, uses a fixed set of vendors, and sits behind regulation, and an organisation that only talks to itself convinces itself that what it does is excellent. Outside disruption is nonetheless hard, because access to infrastructure, the payment rails and the customer's own data, is essential and gated. That single sentence explains much of what follows in this course, and it is why open banking is the catalyst the module dwells on.

v3.0? The question mark is deliberate

The catalyst repeats. What is being automated this time is not execution, which every earlier wave automated, including the high-frequency traders who follow algorithms someone wrote, but decision-making. A mortgage-versus-TFSA question put to a consumer chatbot came back in seconds with the trade-offs, the clarifying questions a good advisor would ask, and no charge; and it will answer the question you would be embarrassed to ask a person. Robo-advice existed for a decade without taking off; what revived it, as the room noted, is generative AI turning "buy this ETF" into something closer to advice.

Which is where the module ends: if your agent is software, whose agent is it?

The links that hold the module together

These are the connections the slides imply but do not spell out. They are why the examples come in the order they do, and they are what the rest of the course is built on.

1. The pattern repeats

A scarce skill or asset is the barrier. A class of people lives on it. A technology makes it cheap. That class, and the authority that ran through it, loses its lever. Coined money let strangers transact without someone to weigh and assay the metal. The crossbow made a peasant a threat to a knight. The printing press took the reading of the book away from the clergy. The bicycle and the car ended the horse economy. Then finance, three times: the telegraph, the smartphone, and now the analyst in your pocket. That is why the section is called the process itself is nothing new. When you meet a new fintech, the first useful question is which barrier is it removing, and who was living on it?

2. Removing one barrier tends to install another

Coined money is the example to keep. It removed the intermediary who weighed the metal, and it created the mint, seigniorage, and a concentration of monetary power in the state that persists today. Intermediation was not eliminated; it moved. Hold this for stablecoins, where the pitch is that you no longer need a bank to hold and move value, and the immediate question is who you need instead, and what they get out of it. The answer, in Canada, involves an issuer holding reserves and the Bank of Canada supervising them, which is a different arrangement rather than an absent one.

3. Incumbents regulate, and the rules have carve-outs

The 1139 crossbow ban came from a Church council in a world run by the armoured nobility the crossbow threatened, and it applied only to use against Christians. Read as economics, it looks like a rule protecting incumbents, with a carve-out, and it did not hold. That is a lens for the whole course. A country with five large banks, none of them losing customers, produces very little lobbying pressure for rules that would make customers easy to lose, which is part of why consumer-driven banking was promised for 2023 and is still not in force. Inertia is partly engineered, too: slow account transfers and exit fees are a rule in that exact form. Neither of these is a conspiracy; both are incentives. Every time you meet a rule in this course, ask who benefits from its existing in that exact form.

4. Catalysts and shocks, and the difference between invention and adoption

This is the idea worth the most in the rest of the course, and it is easy to under-read.

The bicycle's components existed well before 1817. What did not exist was a reason to want one, because horses did the job. Then a volcano erupted in 1815, the following year had no summer, the oat crop failed and horses died or were slaughtered, and suddenly a machine that moved you without a horse had a market. The technology existed; the shock created the demand. The counterfactual is worth sitting with: would the bicycle have spread as far, as fast, if horses had stayed cheap and plentiful? Probably not.

So a shock does two distinguishable things. It can prompt invention, and it can drive adoption of something already invented, and the second is more common and more often decisive. Repeatedly in this course you will find that the technology was sitting there and something else had to happen first. 2008 invented nothing; it destroyed trust and made people willing to try alternatives. The smartphone was not built for finance; it became the channel finance had to move onto. India's demonetisation, raised from the floor, invented no payment technology; it removed the alternative overnight, and digital payments took off. The pandemic invented neither e-commerce nor online banking; it made both unavoidable, and exposed how far behind some systems were.

Two practical consequences. First, being technically right is not the same as being right then: a good idea arriving without a catalyst tends to sit unused, and being in the right place at the right time is something to look for rather than luck to shrug at. Second, when you assess a fintech, look for its shock. If you cannot name what changed to make people willing to switch, be suspicious of the adoption story.

5. Historical path matters, and it constrains what can be built now

This is why the stock-trade diagram is in the deck, and the detail is beside the point. Between Alice selling and Bob buying stand two brokers, a technology provider, an exchange that only matches, a clearing house, a custodian on each side, a separate record of who owns what so the issuer can pay a dividend, and a payment system to move the cash. Each box solved a real problem at a particular moment: custodians exist because nobody wanted their paper certificates mixed with a broker's own. The next piece was added on top rather than in place of the last. Nobody in their right mind would design this from scratch. And we now have a technology that could, in principle, hold all of it on one infrastructure, which is the blockchain conversation, and it comes later. The point of this slide is only to establish that the mess is real and inherited.

Which is exactly why the railroad tale follows it. The story that US track gauge traces back to Roman chariot ruts is debunked, and it is on the slide anyway because it is almost believable: historical accident really does lock in present design. Path dependence is real and expensive, and regulation inherits it worst of all, because a new thing gets layered on the old rule rather than replacing it. The honest version, told as a curiosity, is better still: decades of incompatible gauges, and then, after long preparation, the southern railroads converted roughly 13,000 miles of track in two days in 1886. Coordination that looks impossible for decades can happen very quickly once it is prepared.

6. "Same activity, same risk, same regulatory outcome" is right and hard

IOSCO's principle is technology neutrality: regulate the activity, not the tool. Sensible, and harder than it sounds, because new technology does not leave the risk map intact. A large share of securities regulation exists to police the conflict between you and your broker, who would rather pay the exchange the smallest fee than get you the best price. Send value from your own wallet to a venue and there is no broker for those rules to attach to. Meanwhile something new appears: an exchange ordinarily does not hold your assets, because custody and settlement sit elsewhere, but to trade on a centralised crypto exchange you must first move your assets into its wallet, so it can be broker, custodian and venue at once, and when it fails customers can lose what they held there. That is FTX, and QuadrigaCX in Canada. And the room's question about crime on blockchains got the same shape of answer: the technology removes the broker and adds traceability and a know-your-customer choke point at the on-ramp, while borderless instant transfer lets old crime run at a new scale. Same activity, different risks, in both directions.

7. And the question underneath all of it: who captures the rents?

Start from the cost of intermediation refusing to fall for 130 years. Every wave since has promised to bring it down. v1.0 made markets more efficient. v2.0 built better front ends onto much the same plumbing. v3.0 makes expert advice nearly free, and the early evidence points both ways: AI raises productivity and narrows the gap between an expert and a novice, and AI market makers learn to keep a markup rather than compete it away. Finance workers' rents can go either way depending on the job.

And the sharper version, which is where the module actually ends. The free answer to the mortgage question came from an agent nobody was paying. Would your bank give the same answer? Would an AI agent, if the bank were paying for it? In the Hugging Face incident, AI agents that were supposed to be isolated built themselves a message board out of file names and then directory names, cooperated on a benchmark, and broke into a real company's servers; one agent stopped itself, another posted "go" with a six-minute deadline, and the first took a peer's message as authorisation. Cooperation is what makes agents useful, and it is what made that possible. Your agent took orders from another agent. Whose agent is it?

Open banking may soon let you hand an agent your full transaction history. If it then chooses where you save, where you borrow and how you pay, do you still have a bank, or does the agent pick a different one for each job? That is the app-store structure arriving in finance: the platform captures the customer and everything underneath becomes an interchangeable supplier. BlackBerry did not lose because its phone got worse; what people had actually paid for was mobile email and freedom from the desk, and a platform gave them that with everything else on top. Strip a bank's list the same way and what survives is somewhere for your salary to land and a way to move money. Ask which one the bank is.