Long-term effects of reducing friction costs
Passenger travel and payments
Much of China’s economic growth story has been the story of reducing friction costs. Recently I’ve been looking at two examples of this
Passenger travel — Should China continue to invest in high-speed rail? (also Twitter)
Payments — Ant Group’s upcoming IPO:
There are immediate benefits to reducing friction costs, but there is also a long-term term component that people sometimes forget or under-estimate. Impact from the long-term component is harder to predict but it can easily dwarf economic returns from the direct impact.
High-speed rail — hard infrastructure
With hard infrastructure like high-speed rail, there is the direct and immediate impact of cutting travel time by upwards of two-thirds while also increasing capacity. At first, the direct impact is felt as existing travelers are able to save time and travel in more comfort.
Then it begins to absorb travel demand that had previously been supply-constrained. For example, in China passenger rail was notoriously capacity constrained especially during peak demand like the Lunar New Year migration. As capacity increased, there are fewer of these stories today.
Next, travel demand that was previously impractical is unlocked. While travelling eight hours may have been too inconvenient to justify for a leisure trip, cutting the time to three hours may change the cost-benefit analysis. As anyone who’s taken a 16-hour flight before knows, the marginal discomfort of the last half of the trip is sigificantly higher than the first half of the trip. Domestic tourism has been a large beneficiary of HSR.
Finally, the long-term second- and third-order impacts are based on people, businesses and systems consciously making changes to their existing lifestyle to adapt to this new and improved infrastructure that is now available. As they figure out new economic opportunities, they may move or re-locate businesses to take advantage of improved transit options.
In the beginning, the economic returns might be muted (or even depressed, as economic resources are allocated to long-term oriented investments) but in the long run as society naturally optimizes and adapts, the dividends start trickling in. After more time passes, the trickle may turn into a torrent.
Agglomeration effects
Geoffrey West draws some interesting parallels between the complex biological organisms and complex social organizations like cities.
The ideas in Scale revolve around the idea that just like you can predict certain characteristics of biological organizations based on their size, you can also predict certain characteristics about cities based on their size, signalling some natural realtionship between the two.
One really interesting finding was how the productivity of cities increased super-linearly with the size of the city. In other words, as cities got bigger, they generated more GDP not just because they are bigger, but because their individual units were also more productive.
Transportation systems feature heavily because they are key enablers of physical size.
Without the subway, New York City could not have physically supported the population densities we saw by the first half of the 20th century.
High-speed rail enables the potential for ultra-large megalopolises. First, there were towns. Towns became cities. Cities became megalopolises. And now now we may be on the verge of witnessing the next level up.
China wants to combine existing mega-cities like Beijing with their surrounding areas to turn them into even bigger versions of themselves. HSR solves some of the practical limitations — it is hard to be a single integrated city if it takes more than two hours of travel.
Shinkansen helped enable Tokyo-Yokohoma to grow into the largest megalopolis in the world, with 38 million residents. With trains that are effectively 30-40% faster, China is trying to create ultra-large megalopolises with upwards of 100 million people.
If Professor West’s thesis holds true, then we would see yet another leap in relative productivity that comes with getting to this threshold. This is a long-term effort measured in decades, going back to my original point about long-term dividends but the impact would be enormous. [1]
Payments — soft intrastructure
Before Alipay came along, China was effectively an all-cash economy. Pluck someone out of Tang-era China, hand them a wad of cash and they would have had few issues navigating a traditional wet market.
Smartphone proliferation was the critical enabling factor for digital payments. Now any phone could be turned into a payments terminal and start receiving payments.
It’s not a surprise that credit cards never really took off in China — they were leapfrogged by mobile payments.
For merchants, accepting digital payments through your phone doesn’t actually save you money at the point-of-sale. Indeed for certain aspects, it costs more — as shown in this chart, the estimated direct cost of processing cash is lower in some cases than merchant fees on Alipay/Wechat (at least back in 2017):
The biggest reduction in friction cost is related to trust. Using cash involves trust. Trust that the wad of cash handed to you is correct. Trust that the cash is not counterfeit. Trust that your employees won’t steal it.
These costs add up — 5-15% of revenue depending on retail category based on recent studies.
Outsourcing trust to a third-party intermediary like Alipay reduces this friction cost to almost nil. The merchant fee that is paid to Alipay captures the cost of authentication while providing plenty of room for Ant to generate a profit.
Similar to high-speed rail, the immediate beneficiaries were existing merchants. Then it was new merchants that expanded or entrepreneurs that started new businesses to soak up increasing demand. Then it was entirely new industries that were created on the back of new ways of paying people (e.g. livestreaming) or purchasing micro-services (e.g. dockless bike-sharing).
Again, these are just the direct impacts of reducing friction costs of payments.
By moving to digital payments, merchants started collecting data in systemic and structured way that just wasn’t possible in a cash-based world. This data enables merchants to make more informed decisions about their individual business — when to replenish inventory, whether to invest in a piece of new equipment, whether to expand into a new location, which marketing channel to spend money in, etc.
Pooled together and packaged by intermediaries like Ant, this data can generate actionable insights into the broad economy — improved credit scoring algorithms, better investment decisions, more efficient allocation and deployment of resources.
Economic development is a continuous optimization process that is based on individuals and businesses trying to separate signals from noise to make decisions. Structured transaction data is one of the most powerful signals available because it is based on pure consumer intent.
Just as with high-speed rail, it is the second- and third-order effects that will really matter and make an impact in the long run. As I wrote earlier, Civilization’s Research Tree is a helpful framework to think about innovation and its second- and third-order (and beyond) effects.
Using this framework, one can’t help but wonder about the long-term impact of 100-million person cities with lightning-fast signal propagation.
Note
[1] This is not a prediction that Jing-jin-ji will be more productive than Greater NYC MSA or even Tokyo-Yokohoma. The super-linear productivity scaling relationship works within the confines of an economic zone, which can differ in productivity.
This was originally published on Twitter in August 2020.