Friday, January 26, 2018

Bitcoin and financial bubbles



Great Economist Buttonwood article on Bitcoin and financial bubbles:

"Perhaps the best way of understanding bitcoin is through a model of how bubbles operate. The classic model, developed by Hyman Minsky and elaborated by Charles Kindleberger, a historian who studied bubbles, has five stages: (1) displacement, (2) boom, (3) euphoria, (4) financial distress and (5) revulsion. The displacement is some technological development that can be used to justify a “new era”—railways, canals, the internet or blockchain. A boom then occurs and drags in more and more investors; at some stage, we reach euphoria, where the boom is widely known to the public and there is talk about those who made millions from the trade.

In the euphoria stage, people buy because others are buying and because they anticipate being able to sell quickly at a higher price. For a while, this trend is self-reinforcing. At some stage, however, doubts set in; some people decide to take their profits while they can. Bad news occurs; with bitcoin, it seems to be reports that authorities in South Korea (where trading has been particularly active) are going to crack down.

Once the price starts to fall, the psychology changes. People who bought early and were counting their millions suddenly see a dent in their wealth (and it is worth noting that you are not really rich unless you have got into the asset class and out again). Other people may have bought above the current price and are bitterly regretting their mistake. Bargain hunters jump in and temporarily drive the price higher but it doesn’t last.

We have not yet reached the “distress” stage but we might be near it. Worries about the security of cryptocurrencies could be the trigger for another sell-off.  Investors may well reflect that bitcoin had not become a means of exchange for day-to-day transactions, has not proved to be a reliable store of value and thanks to the proliferation of cryptocurrencies, does not really benefit from a limited supply. [...]
A digital asset that has no income stream is very hard to value. That makes it hard to designate a price target on the way up, but also hard to set a floor on the way down. But by the time people realise that, we will be in the “revulsion” stage."

Now is Bitcoin a bubble?  There are many arguments for "Yes", including:

  • Bitcoin is NOT blockchain, which is useful and is being adopted by many financial institutions in novel ways.  There are iterations to the Bitcoin idea that can evolve, and there's really zero barrier to entry i.e. a new iteration can just replace Bitcoin as the dominant cryptocurrency.
  • Bitcoin is not usable as a currency, since there's a limited number of transactions per second that can be processed, and the volatility of its value means people would rather not trade Bitcoins (and people who traded with Bitcoins in the past now regret it).  Indeed, seems like nobody really knows what Bitcoin should be used for nowadays.
  • Bitcoin network takes way too much energy to maintain.

I'll leave you with Ronnie Chieng:


Wednesday, January 17, 2018

North Korea nuclear crisis: Lessons from History


The possibility of a 21st century nuclear holocaust, pitting a pudgy dictator against the septuagenarian wanna-be-authoritarian leader with a (ahem) bigger nuclear button on the other side of the Pacific, is real enough to haunt our worst nightmares.  Last weekend's false alarm in Hawaii made the threat particularly close to home for the island's 1.4m inhabitants.

A fascinating well-written article by George Perkovich chronicles the thought processes during the 1962 Cuban Missile Crisis: what President John F Kennedy and Defense Secretary Robert McNamara understood and what they didn't at the time, and the roles of Soviet premier Nikita Khruschev and Cuban leader Fidel Castro.  The uncertainties, misjudgments and miscalculations, could have lead parties astray towards escalation.  Historians James Blight and Janet Lang went even further, saying the reality was far more dire than the parties understood at the time: "if the crisis were run under the same conditions 100 times, it would end in nuclear war 95 times. We are living in one of the five alternate universes in which humanity survived."

Definitely within missile range

Most of us know the story:  during the height of the cold war, US spy planes detected Soviet ships mobilizing missiles and nuclear materials into Cuba -- less than 100 miles off the coast of Florida.  JFK responded by installing nuclear weapons to the US base in Turkey, right on the Russian doorsteps.  Then an American U-2 spy plane was shot down by Cuban anti-aircraft guns, with several others escaping intense enemy fire. Military leaders pressed the president to authorize an invasion.  JFK held still, and sent his attorney general brother Robert "RFK" Kennedy to a secret meeting with the Soviets.  Lengthy negotiations eventually lead to both parties standing down whereby Khruschev agreed to withdraw the Cuban missiles  and the US withdrew its own from Italy and Turkey  -- the American public were not even privy of the negotiation details.

What nobody -- not the Kennedys, not McNamara -- knew until recently, was terrifying, to say the least:
  1. Americans knew the Soviets were mobilizing weapons into Cuba, expecting  eventually a handful intermediate-range ballistic missiles able to reach continental United States.  But nobody knew the Soviets had already deployed more than 100 battlefield-ready tactical nuclear weapons in the island, and that local Soviet officers were pre-authorized to use them in the widely-expected Marine invasion.
  2. Furthermore, McNamara told Kennedy at the time that he believed there were 8,000-10,000 Soviet military personnel in Cuba.  In 2008 it was revealed that the actual number was far higher at 42,000.  Had JFK agreed with his generals to invade Cuba, the landing and subsequent ground war would have been unexpectedly lengthier, perhaps taking in much larger casualties, which may lead both sides to turn to their advanced weaponry. 
  3. On Oct 27, 1962 during the blockade of Cuba by the Americans, a Navy warship dropped depth charges on a Soviet submarine.  Years later it was discovered that the submarine carried nuclear torpedoes, and it was out of radio contact with the Kremlin and believed that war has started.  Two Soviet officers on board, including the captain, wanted to fire the torpedoes, but one refused -- and catastrophic war was narrowly avoided.
  4. Finally, US envoy RFK pressed his Soviet counterpart to stop firing at American military planes.  What he didn't know is that Khruschev never authorized such attacks, but Castro did.  The Cuban dictator feared an imminent invasion, and would rather risk total annihilation over American troop presence in his island state.  The Soviets agreed to RFK's demands, but had Castro kept his aggressive stance, this could have lead to mutual distrust and escalation. 


It's important to note that we probably understand less about North Korea now, than we did about the Soviets back then. The possibility of miscalculations are enormous, the consequences apocalyptic. Some of these fallacies may include:
 
  1. Misjudging the effectiveness of our missile defense system.  Technology has advanced since the Cold War, but these batteries have only been tested in isolated exercises.  We don't know exactly how effective they would be in real-life war situations. Misplaced confidence on the defensive measures can lead to bad decisions. 
  2. President Trump believes that China can coax North Korea to drop its nuclear ambitions. However, despite China's large economic contribution to North Korea, Kim Jong Un has pared down China's direct influence in his country.  Pro-Chinese elements in his political circle have been purged, most notably his brother Kim Jong Nam, who was assassinated (widely believed to have been perpetrated by North Korean agents) while in Hong Kong's protective asylum. 
  3. Grassroot conditions are also a big question mark. We hear about massive drought affecting North Korean food supply, and crippling economic sanctions bringing about the prospect of a humanitarian crisis.  But North Koreans are indoctrinated since early childhood; they learn to shoot rifles the same time they learn to read and write; they spy on their peers to identify moles and traitors; they venerate the Kim dynasty like a deity; and they despise America like the devil.  Pentagon puts the number of troops under Kim at about one million, but determination may end up meaning more than sheer numbers.  Just ask Americans about the Vietnam War. 
  4. Defense Secretary Mattis proposes a limited strike on North Korean missile sites, just to give them a "bloody nose" and show them we are serious. However, this is a fallacy at best -- how can a nuclear-equipped adversary understand the attack's "limited" nature? During the cold war, the US military scrapped its low-yield strategic nuclear weapons program, because (i) with tensions running high, it's difficult for Russia to see the difference between "low yield" and "normal yield"; and (ii) even if they can, Russia may not be able to respond with a similarly-limited strike.  Similarly in North Korea, Kim may see any attack on his country as an all-out existential threat worthy of highest possible retaliatory escalation.
What's the best policy recommendation in this case? I don't know, but the best argument I've read is to nuke Kim Jong Un's presidential potty.  Now that's thinking outside the box! 

Thursday, January 11, 2018

How NOT to write an astronomy article

Oh dear....

Linked article here.  Beware of "has Uranus been probed" jokes.

Tuesday, January 02, 2018

What's the endgame for Ride-Sharing?

Uber's non-dominance in SEA markets

2017 was the year of reckoning for Uber due to a barrage of scandals, culminating in a massive coup/down-round investment by Japanese investment firm SoftBank. But it's worth taking a closer look at the nature of the business itself.  Ride-sharing (also known as ride-hailing, or on-demand mobility) companies like Uber and Lyft are designed to be an intermediary business: they have no assets, take on no risk, no nothing, just technology-based platform for the transport industry, connecting riders and part-time drivers wanting to make additional income.  In Southeast Asia, the market leaders are Go-Jek and Grab, with a slightly different business model of enrolling professional drivers, helping them with car ownership/rentals, and providing access to the app.  How any of that justifies billion dollar valuations, you tell me.  One may argue that an asset-heavy business should justify bigger valuation, but recent crashes in Chinese bike-sharing startups (due to either poor asset management, or mismanagement in general) should put that theory to rest.

So this begs the question: what's the real endgame for ride-sharing? The companies simply can't survive by doing this forever, as both drivers and riders are spoiled by subsidized economics. Is it logistics? Food delivery? Fintech (e.g. do they want to end up like Alipay)? If Uber fares rise significantly to levels that make the business sustainable without subsidies, won't the market also shrink greatly (e.g. people would just return to owning cars or, gasp, taking public transport) ? Would self-driving cars be the endgame? How do we even know that driverless Uber taxis make a good business?

Let's explore.

Rides are fundamentally unprofitable

Transportation industry analyst Huber Horan lays out the case for skepticism over Uber's hype. Amazon saved costs by getting rid of expensive retail stores and sales staff.  But Uber rides still require a car, a driver, and fuel -- just like any old taxi rides.


OK then, many taxi companies were making decent profits for some time before Uber; wouldn't ride-hailing be at least as profitable as these old companies?  No, for three reasons, eloquently laid-out by Len Sherman in Forbes:

(a) The taxi industry that Uber is seeking to disrupt was never profitable when allowed to expand in unregulated markets, reflecting the industry’s low barriers to entry, high variable costs, low economies of scale and intense price competition -- and Uber’s current business model doesn’t fundamentally change these structural industry characteristics.

This historical experience exhibits several parallels to Uber’s current business model, presaging the company’s dismal financial performance. The pre-regulated taxi industry was characterized by bounded demand, abundant supply, relatively undifferentiated service quality, extremely low barriers to entry, low customer switching costs, high variable costs and virtually no economies of scale.  [...]

(b) Conventional wisdom has held that Uber would enjoy a global winner-take-most outcome because of their outsized balance sheet and strong network effects. But to the extent that network effects exist, they are local, not global. For example, Uber's 89% market share in Tampa doesn’t help them in Portland, where Uber’s market share is below 50%. In fact, Uber has struggled to achieve market share leadership in many large foreign markets, including China, India, SE Asia and Brazil. Moreover, while network effects do exist within each metro market, the benefits are significantly weakened by extremely low switching costs, which enable drivers and riders to utilize whichever ridesharing service offers the best deal on any given trip. [...]

(c) [There's] an inherent conflict between the business objectives of Uber and its drivers. Uber’s revenues are directly proportional to the number of trips it can facilitate, and thus the company has strong incentives to continuously scale its business. Drivers of course want to maximize their revenue per hour worked. But as Uber continues to recruit drivers, the revenue potential per driver inevitably declines. As the highest revenue-generating neighborhoods become increasingly saturated, new drivers are forced to seek less attractive service territories to find customers.



Ride economics are heavily distorted

Although we pay less for Uber rides than taxis, our views are highly distorted; it just seems like rides are cheaper because the company is making a loss on every ride, with a dream of dominating each market it enters  Uber burns cash at US$2bn per year globally, mainly on subsidies and new market expansion.   Similarly in Indonesia, leaked investor presentation shows market leader Go-Jek spent US$73mm on subsidies in the six months between Oct15-Mar16, yielding 12% growth in completed orders over that period.  For that period, subsidies actually made up 72% of the company's gross booking fare value.  The distortion affects the economics for consumers as well as drivers.


Challenging economics

Not just subsidies, gamification also distorts the psychology, giving both riders and drivers real and false rewards for spending/loyalty to the platform.  Nightmare stories and complaints about exploitation and lack of transparency abound -- especially as the platforms naturally always side with riders over drivers.

Ride-sharing creates serious public policy issues

Setting aside the debate over independent contractors vs employees (which have not been settled), or the corporate culture/moral compass issues, or the whole evading the law thing, recent analysis from extensive data in New York City shows that ride-hailing apps are significantly worsening the traffic problem, while also impacting the ridership and quality of public transport.  “For years, as the city grew, more and more people took the subway and bus,” according to Bruce Schaller, a former NYC DOT official. “Now, as the city grows, more and more people are taking Uber, Lyft, and Via. This is not a sustainable way for the city to grow.”  He continues by noting the drop in quality of New York City’s mass transit. Not only has MTA bus ridership rapidly dropped in recent years, but the subways are also now losing passengers, too. Service quality deteriorates, and delays and mechanical failures have become the norm rather than the exception.  Clearly we still don't have a grasp of the real, long-term public policy impact of ride-sharing.
The picture in Singapore is no different

The endgame remains a question mark

SE Asian market leaders are taking a stance: Go-Jek and Grab wants to expand beyond ride-hailing, and into fintech through organic growth and acquisitions.  Grab CEO Anthony Tan said the company wants to “win payments in Southeast Asia,” saying that the app’s user base can be the groundwork for more services.

Robots are employees or independent contractors?
Uber's endgame is less clear.  Are they just buying time until things figure themselves out? After all, Amazon spent many years and many billion dollars in shareholders' equity, evolving from selling books to selling CDs to selling everything to producing own hardware to producing own content to selling cloud services to selling groceries (did I miss anything? something with drones?) -- all in the hunt for that elusive profitability...