Debunking the Theory of the Firm

Until recently, US tech companies were pretty good about taking care of employees from new hire to retirement. Many Fortune 500 companies had explicit no-layoff policies: Hewlett Packard, Motorola, General Motors, McDonnell Douglas, Lincoln Electric, American Airlines, Delta. IBM never laid off a single worker until 1993.

This tie clip is a tiny slide rule that IBM gave to retiring employees. Do they still give these out? Do employees even make it to retirement age anymore?

At some point, the employer-employee relationship fell off a cliff. Corporations used to value the loyalty they gained by promising lifelong job security. Now they don’t even want real employees: Nearly all of the 10 million jobs created since 2005 are temp positions.

Does this disprove the Theory of the Firm? According to Ronald Coase, organizations form long-term relationships with employees to eliminate the transaction costs of constant market exchange. Sourcing candidates, negotiation, hiring with incomplete information, making sure contractors don’t run off with a USB stick full of trade secrets – that’s all really expensive!

The Sovereign Individual predicted that technology would eventually automate the firm away. Information systems and AI could seamlessly coordinate a two-sided marketplace. Offices equipped with surveillance devices would measure workers’ output, obviating the need for employee trust. Isn’t that basically Uber? With the help of services like LinkedIn and Gigster and Foundry and Fiverr, we can already reduce transaction and coordination costs to the point where full-time employment makes no sense at all.

Why stop at ruining jobs? Marriage commitments are similarly passé. Humans used to mate monogamously because fathers had to be reasonably sure of paternity before investing in child-rearing chores. It took years to build up that kind of trust. With DNA tests, guys can quickly determine which kids to care about. Better yet, forget about turning parenting into a joint effort — just outsource it to an on-demand app.

Uber for Airdropped British Nannies.

Just as technology has lowered the marginal benefits of retaining a dedicated mate, it has also reduced the costs of sourcing such mate. We used to waste so much time on courtship and gentlemen callers and other nonsense, but now we can just swipe on Tinder and Grindr and OkCupid and Backpages.

Firms and family units are dead. That numbness you feel is full-blooded individual empowerment.

See Also:
Contracts and Trust

Go Away Amazon

In the middle of a self-declared housing emergency, the San Francisco Bay Area Council has respectfully submitted this bid inviting Amazon to build HQ2 here.

NOoOOoooooOOOOOO!! Why would you do that??

We already have Google, Facebook, Apple, Netflix, Uber, Lyft, Tesla, Twitter, and about 99% of the Russell 2000. Are we trying to make the rest of the country resent us more than they already do? Because I don’t think that’s possible.

Amazon promises to create 50,000 new jobs. We don’t need more jobs! Our unemployment rate is so low, it’s negative. Like, people are commuting from three hours away just to work here. Heck, our unemployment rate is so low, even homeless people have full-time jobs.

(Or maybe our housing situation is so messed up that even full-time workers can’t afford the rent here.)

Here’s an artist’s rendering of the proposed site of HQ2:

The conceptual rendering looks lovely from afar, but if you zoom in to the street level, you might see something like this:

Right. San Francisco’s suggested location for HQ2 is Hunter’s Point, home of the largest homeless population in the city. Maybe Mayor Lee hopes that Amazon will pave right over them?

To be fair, San Francisco isn’t the only site offered in the bid. The Bay Area Council listed all the Northern Arc cities – basically everything connected by BART. Today, the East Bay BART cities represent the last bastion of affordable housing in the Bay. In Concord, a 1.5-hour commute from the city, the average 2-BR apartment rents for just under $2000 a month. Richmond, a mere 75 minutes away, rents two-bedrooms for $2500. Expect these prices to double once Amazon moves in.

I can’t blame the council for this submission, especially if it means Bay Area residents will get their Amazon Prime crap delivered in under an hour. Also, there’s potentially a huge chunk of tax revenue on the line. Who knows how much — it’s large enough that Governor Brown felt justified in offering hundreds of millions in tax breaks to incentivize an Amazon HQ in California.

Although my post is selfishly motivated, I do think that everyone will be better off if Amazon goes elsewhere. There are 237 other bids out there; undoubtedly some other metropolitan area needs the jobs more than we do. Mayor Lee, back off and let Detroit have this one.

This is the last page of the Amazon bid. The eggs represent the shattered dreams of Bay Area millennials who thought they might have a shot at home ownership in their lifetime. Hahaha, nope!

The Transaction Costs of Tokenizing Everything

I wonder if Al Gore ever looks down at us peons, crawling around the internet like eight-legged leeches:

I invented that. I took the initiative in creating the Internet. Now all these freeloaders are using MY internet protocol to drive billions of dollars worth of value. For FREE.

Damn, I should have done an ICO.

Even though Al Gore neglected to tokenize his internet protocol*, someone else came along with the next-best thing.

In 1999, a clever company called Enron invented something called a bandwidth contract.

The internet is just a bunch of routers and cables, sending and receiving data all day long. Most internet providers have peering agreements, where they carry each other’s traffic for free. Sharing is mutually beneficial, and their customers pay a fixed monthly rate regardless of use.

That’s all well and good when capacity is plentiful, but what happens if half the country wants to stream Sunday Night Football while I’m trying to sync my Bitcoin node? Whose data gets to go first?

Enron’s bandwidth contracts were designed to solve this potential queueing problem. By forcing internet users to bid for bandwidth by the minute, the free market would decide the optimal allocation of resources [1].

Sadly, Enron imploded before it could fully realize its bandwidth trading dream. Still, the idea of turning every network into a market was pretty hot in the dot-com days [2]. To see how things might have turned out, we can look at a company called Mojo Nation.

A MASSIVE AMOUNT OF STORAGE SITS UNUSED IN DATA CENTERS AND HARD DRIVES AROUND THE WORLD. Let your hard drive shit out money by fulfilling storage requests on the open market!

Such is the marketing pitch of services like Filecoin, Sia, Storj, MaidSafe, and all those other decentralized file storage tokens. Seventeen years ago, their founders were still in diapers when Mojo Nation launched to address the problem of Pareto-inefficient data storage.

Mojo Nation created a digital payment system to buy and sell computational resources. Participants could earn Mojo tokens by contributing things like disk space, bandwidth, CPU cycles. Those who wanted resources offered bids in outgoing requests. Mojo tokens relied on a centralized mint because blockchains weren’t around yet, but centralization was the least of its problems: Tokens were a huge distraction from what users really wanted to do, which was share files [3].

A bidding market is an awfully complicated thing. Take Bitcoin, for instance. Each block has a finite capacity, so participants submit transaction fees to incentivize miners to include their transactions. It’s a simple concept, but transaction fees are the most aggravating part of Bitcoin. There are people like Roger Ver who have been using Bitcoin since 2011 and STILL can’t figure out how transaction fees work.

I’m not trying to pick on Roger here; this is not a user-friendly experience. Those who want to tokenize all the protocols are effectively shoehorning the same shitty experience into every aspect of the internet. My mother can’t even update her Facebook picture without backup assist; how on earth will she manage five-dozen protocol tokens to navigate the web?

Many dot-com era platforms tried to create bandwidth exchanges, but none found willing participants. Enron and Blockbuster temporarily joined forces to create on-demand streaming video, in hopes that they could clog up so much bandwidth that internet providers would start a bandwidth bidding war. No such luck. As it turns out, bandwidth — and most computational resources — are simply too cheap to meter.

After Mojo Nation’s demise, a former employee stripped the token incentives out of the protocol and created a simple tit-for-tat filesharing system. The software client uploads files to peers that provide downloads [4]. Users can’t accumulate credits, and sometimes freeloaders go unpunished, but people don’t care about perfect resource allocation — they just want convenient file access. By 2004, BitTorrent was responsible for a quarter of all the traffic on the internet.

And everyone lived Pareto sub-Optimally ever after.

* Kidding. The Internet Protocol was created by Vint Cerf and Bob Kahn. The only thing Al Gore invented was global warming.

1. Enron’s Bandwidth Trading patent, 2001.

2. Mark Miller and K. Eric Drexler. The Agoric Papers, 2000.

3. A conversation about Mojo Nation on Unenumerated, ca. 2007.

4. Bram Cohen. Incentives Build Robustness in BitTorrent, 2003.

5. Bryce Wilcox-O’Hearn, who now goes by Zooko, CEO of ZCash. Mojo Nation: Experiences Deploying a Large-Scale Emergent Network, IPTPS 2002.

6. Mojo Nation website from 2000.

If No One Spends Bitcoin, How Can It Have Value?

Medieval mint, engraving by Leonard Beck (1516).

It’s hard to imagine a world without penny candy and nickel newsreels, but for most of human history, petty transactions were a pain in the ass.

Prior to the Industrial Revolution, coinage was a labor-intensive process. Metal had to be melted, refined, hammered, and cut. Because it took just as much effort to hammer out a small coin as it did a big one, mintmasters were inclined to create only the largest denomination coins.

If it weren’t for taxation and church collections, the state would have had no reason to issue small denominations at all. To encourage the creation of small change, medieval states authorized seigniorage — mints reduced the relative quantity of silver in small denominations to offset production costs.

Production costs of coinage (brassage).

Debasement! Where legal tender laws are enforced, bad money drives out good. Creditors complained that debts were being repaid in shittier coins than what was lent out. In states without legal tender, the large-denomination coins became the unit of account, and smaller coins had a floating exchange rate depending on their level of debasement.

The more the small denominations were debased, the worse the exchange rate got. Seeing small denominations as a poor store of value, people melted them for the commodity silver, exacerbating the small-change shortage. Small coins provided liquidity, but the liquidity service was not valuable enough to counteract debasement.

The biggest transaction cost is trust.

The title question is backwards. Value does not come from the ability to spend; the ability to spend comes from value. The full-bodied large coins were more valuable than liquidity-providing small coins because the gold and silver content securely constrained their supply.

There are plenty of cheap solutions for illiquidity. When small coins were scarce, retailers and craftsmen issued lead and copper tokens as a substitute for change. The tokens had no commodity value, but customers accepted them because they trusted their local businesses.

Foreign trade doesn’t have the benefit of localized trust, so merchants must rely on a scarce and unforgeable intermediate commodity. The Group 11 elements (Copper, Silver, Gold) have been universally employed as coinage metals thanks to the eons-old neutron star collisions that created a limited supply. Their shared electron configurations make them pliable and corrosion resistant. Atomic weight corresponds to the required energy input and hence, unforgeable value.

There’s no cheap substitute for securely constrained scarcity, especially since humans are so good at making scarce things abundant. Domesticated livestock, designer knockoffs, genetically engineered plants. Even labgrown diamonds are nearly indistinguishable from the real thing. We try to mimic scarcity with patents and licensing and zoning regulations, but these are all expensive solutions. Fiat money pretends to be scarce, and that costs us $600 billion a year.

People are trying to make Bitcoin knockoffs. After all, it’s human nature to want to make a scarce thing abundant. If they succeed, then Bitcoin has no value. But if Bitcoin has no value, how can anyone spend it?

See Also:
1. Nick Szabo, Unforgeable Costliness, 2004.
2. Thomas J. Sargent & François R. Velde. The Big Problem of Small Change, 2002.