Austin Contrarian

Unplanned Cities (paper)

This paper by law professor David Schleicher (George Mason) is pertinent to my argument against SMDs:

Generations of scholarship on the political economy of zoning have tried to explain a world in which tony suburbs run by effective homeowner lobbies use zoning to keep out development, but big cities allow relatively untrammeled growth because of the political influence of developers. Further, this literature has assumed that, while zoning restrictions can cause "micro-misallocations" inside a metropolitan region, they cannot increase housing prices throughout a region because some of the many local governments in a region will allow development. But these theories have been overtaken by events. Over the past few decades, land use restrictions have driven up housing prices in the nation's richest and most productive regions, resulting in massive changes in where in America people live and reducing the growth rate of the economy. Further, as demand to live in them has increased, many of the nation's biggest cities have become responsible for substantial limits on development. Although developers are, in fact, among the most important players in city politics, we have not seen enough growth in the housing supply in many cities to keep prices from skyrocketing. 

This paper seeks explain these changes with a story about big city land use that places the legal regime governing land use decisions at its center. Using the tools of positive political theory, I argue that, in the absence of strong local political parties, land use law sets the voting order in local legislatures, determining policy from potentially cycling preferences. Specifically, these laws create a peculiar procedure, a form of seriatim decision-making in which the intense preferences of local residents opposed to re-zonings are privileged against more weakly-held citywide preferences for an increased housing supply. Without a party leadership to organize deals and whip votes, legislatures cannot easily make deals for generally-beneficial legislation stick. Legislators, who may have preferences for building everywhere to not building anywhere, but stronger preferences for stopping construction in their districts, “defect” as a matter of course and building is restricted everywhere. Further, the seriatim nature of local land use procedure results in a large number of "downzonings," or reductions in the ability of landowners to build "as of right", as big developers do not have an incentive to fight these changes. The cost of moving amendments through the land use process means that small developers cannot overcome the burdens imposed by downzonings, thus limiting incremental growth in the housing stock. 

Finally, the paper argues that, as land use procedure is the problem, procedural reform may provide a solution. Land use and international trade have similarly situated interest groups. Trade policy was radically changed, from a highly protectionist regime to a largely free trade one, by the introduction of procedural reforms like the Reciprocal Trade Agreements Act, adjustment assistance, and "safeguards" measures. The paper proposes changes to land use procedures that mimic these reforms. These changes would structure voting order and deal-making in local legislatures in a way that would create support for increases in the urban housing supply.

I'd call this an argument for a weak version of "ward courtesy."  The "deal" he proposes would fail in Austin due to the sizable block of voters who lack even a weak preference for an increased housing supply.

By the way, pages 18-30 is a very nice, readable introduction to the law and economics of zoning.

(Aside: My land-use professor, Robert Ellickson, features prominently in the discussion.  He was one of the first to recognize, way back in the 1970s, that the growth controls just then beginning to flourish in places like California not only make housing unaffordable, they are economically inefficient.)

(Aside to the aside:  I participated in the landlord-tenant clinic one semester, as a law student.  We represented poor tenants in New Haven, almost exclusively in eviction proceedings.  In practice, this meant going through the paperwork and finding small technicalities to force the evicting landlord to reissue notice and recommence the proceedings.  Delay offered the opportunity to extract a concession for our clients.

Thanks to the success of prior generations of students, most experienced landlords hired from a small stable of lawyers in New Haven, who charged through the nose but who didn't make stupid paperwork mistakes.  We could still usually find non-frivolous arguments to warrant a hearing, even if the loss rate was pretty high.

We had a weekly seminar as part of the clinic.  Ellickson came to speak once.  He told us that, while we had good intentions, we were just driving up housing costs for the average low-income tenant in New Haven.  We, a bunch of leftist idealists, were skeptical.  But, in retrospect, he was probably right:  in New Haven circa 1990, the standard security deposit was two full months rent.  That's a lot of liquidity for a renter, and I'm pretty confident it was due to the difficulty of evictions.  There would be a lot more people in Austin looking for housing if that was the standard here.)

 

The Freeway Congestion Paradox

I've gotten some pushback on my claim that adding two congestion-priced lanes to MoPac will increase capacity more than adding two free lanes.

This is the Freeway Congestion Paradox:

Chen and Varaiya, in their article entitled "The Freeway Congestion Paradox," have demonstrated that, once freeway vehicle density (measured in vehicles per mile) exceeds a certain critical number, both vehicle speed and vehicle flow (measured in vehicles per hour) drop precipitously.  They have demonstrated the phonemenon with actual data from a section of westbound I-10 in Los Angeles.  Until 5:10 am, a flow of 2,100 vehicles per lane per hour is maintained, at a speed of 58 mph.  As density increases after 5:10 am, speed steadily drops, until at 7:00 am speed is a stop-and-go 15 mph, and flow decreases to 1,300 vehicles per lane per hour.

The Federal Highway Administration commissioned a study in 2008 that looked at data from Washington, DC freeways.  From the summary:

Highway studies have determined that once traffic volumes exceed the capacity of the roadway, the system can rapidly "break down" to the point where all traffic slows markedly, and the capacity and throughput of the roadway drops precipitously.

They produced neat graphs depicting the drop in capacity caused by stop-and-go traffic, like this one:

This chart pairs vehicle speed and total volume carried (see pp. 15-16 of the report for more discussion).  The dots are connected in temporal sequence, so it's possible to follow the evolution of the traffic jam.  The graph has a a random-walk component, but basically it confirms two things.  First, low throughput volumes are associated with low speed (i.e., stop-and-go traffic).  Second,  after the critical threshhold is passsed, it takes traffic a long time to recover. 

Here is the same chart, except I have drawn over some of the line segments in red.  This shows that once the tipping point is reached, traffic speeds and volumes drop and take a long time to recover.   (Just trace the red line segments starting at "A").

Congestion pricing would have prevented congestion from reaching the tipping point, effectively increasing the highway's capacity.

Rush-hour traffic on MoPac today is the stop-and-go 15 mph type.   Adding a new, free lane would probably shorten the period when the highway is clogged, but it would still be clogged for long stretches of time.  During these periods, a new, free lane would carry less traffic than a tolled lane.

When congestion is not bad, the toll should be low.  That's one of the attractive features of variable congestion pricing -- the price, in fact, should be zero when the marginal cost of congestion is zero.  (According to the Statesman, CTRMA might charge a nominal amount like a quarter to avoid confusing people.  Which confuses me.)  Put differently, the new lane should "act" like a free lane when traffic is not bad.

Congestion, envy and equity

The Central Texas Regional Mobility Authority plans to add two lanes to MoPac between Lady Bird Lake and Parmer Lane.  The cool part of the plan is that it intends to use dynamic congestion pricing -- i.e., the toll will fluctuate as necessary to keep traffic in the the tolled lanes flowing at 50 mph.   The six existing lanes will remain free.

This is how we ought to add new capacity.  It will make everyone better off.   The people who choose to pay the toll will be better off because they value the time savings more than the cost of the toll.  Bus commuters will be better off -- they might be the biggest beneficiaries, in fact -- because they will get a suddenly much shorter commute for (I presume) the same bus fare.  Drivers who continue to use the free lanes will endure slightly less congestion, even if it's just a narrower period of peak congestion.  Finally, taxpayers, if not better off, will be no worse off because they won't have to pay for the extra capacity.  The capacity will be paid by those who value and use it.

Everyone will not only better off compared to the status quo, but . . . and this is the key point . . . they'll be better off than if the two new lanes were free.   Adding two free lanes would reduce congestion, too, but they would reduce congestion less than two new tolled lanes.  Tolling congested lanes increases their capacity.  A congestion-priced lane can handle 1,800-2,000 cars per hour; an unpriced lane during peak congestion will handle less than half of that.  Somewhat counterintuitively, perhaps, charging for a congested lane is a sure way to get more cars through it.

Obviously, drivers happy to pay the tolls to skip congestion would be worse off if the new lanes were free rather than tolled.  Bus riders who would lose out on a faster commute would be worse off if the new lanes were free.  Taxpayers stuck paying for capacity that they might not value would be worse off if the new lanes were free.   But, because tolling lanes increases their capacity, even the drivers who would not use the toll roads anyway would be worse off if the new lanes were free.  They would have to contend with more congestion.   

Congestion pricing is a no-brainer.  So I was puzzled when Ben Wear raised the equity angle in a companion column.   After pointing out that all drivers would benefit from the congestion reduction from the new lanes, he concluded:

Undoubtedly however, those with money will be able to use the variable-price lanes more often than those on a budget.

It may become something of a spectator sport for those caught in MoPac congestion to count the number of Lexuses, Mercedeses and Escalades that fly by in the express lane.

It ain't necessarily fair. But in the words of John F. Kennedy (who certainly could have afforded to use this lane), life is unfair. And it will help pass the time.

It's strange to complain about fairness when everyone will be better off with new tolled lanes than new free lanes.  

But it's especially strange to complain about fairness because, once the lanes open, the only people on MoPac who will be causing a net social loss will be the drivers in the free lanes.

 This is basic congestion theory but most people still don't get it.  When you enter a congested highway, your mere presence inflicts a cost on everyone else.  Every car behind you is delayed just a little bit, maybe only a few seconds each, but the delay affects a bunch of cars.  Getting in everyone else's way imposes a social cost.

There's not a lot a driver can do about that today.  We can't even say for sure what cost each driver is imposing on everyone else.  But once the toll lanes are built, we'll know.  We'll know because the congestion cost each driver imposes on his fellow drivers will be approximately equal to the price of driving in the congestion-free lane.  That's one of the quirks of the economics of congestion -- it turns out that the cost of congestion a driver imposes is equal to the congestion-clearing price.

Once the toll lanes are built, drivers will have a choice.  The people who enter the tolled lane will be precisely the drivers willing to pay the cost of their congestion.  The toll lane will allow them to pay that cost and get out of the way of everyone else.

The people who don't pay the cost are people for whom, say, 15 minutes in time savings isn't worth the $3 toll or whatever it turns out to be.  That's most people, presumably.  But saving 15 minutes is worth something to them.  If you value the time savings at $1, you would incur a $2 loss by taking the tolled lane.  But you will impose a $3 loss on your fellow drivers if you stay in the free lanes, because we know the cost of your congestion is equal to the toll.  Staying in the free lanes produces a net social loss of $1.

Congestion is socially-sanctioned selfishness.  It's ok only because everyone does it.  But if we'd get used to thinking in terms of costs we're inflicting on everyone else, there'd be less moaning about those paying for the cost they're imposing while we mooch our way down the free lanes.  

Cops and water lines

As the comments here and here show, discussions of subsidies for suburban development often become entangled with discussions of the relative cost of city infrastructure and suburban infrastructure, and with density's impact on those costs.

I'd like to focus here solely on the dynamic of infrastructure costs.  The distinctions between capital costs, long-term variable costs and short-term variable costs seem to get muddled in the discussion of infrastructure, so let me use police costs as an analogy.  It will help to make some simplifying assumptions.

Imagine a city hiring its first police force.  It has to pay an upfront cost for training the policemen.  But after it makes that investment, its initial yearly labor expense is pretty low.  Young policemen aren't paid much, their health expenses are low, and their pension payments aren't due for many years.

But as the police force ages, it gets more expensive.  The city has to pay higher salaries.  Older policemen need more medical care.  And, eventually, everyone notices the looming, gigantic pension expenses.  The mayor and city auditor and the local newspaper begin to moan about excessive health and pension benefits and shortsighted politics.

Suppose the city's population is constant.  The residents pay the upfront cost of the police either through initial assessments incorporated into home prices or a tax levy.  But the per capita annual cost of the police force is low at the beginning.  As the police force ages, the per capita cost rises if, as we've assumed, the city maintains a stable population.  If the city's residents are in fact sensitive to their tax burden, though, some will move away rather than pay the higher per capita fees.  This will increase the per capita burden on the remaining citizens, which will encourage more to move, and so on in a vicious cycle.   (Something like this has played out in many aging midwestern industrial cities.)

Allowing density to rise breaks this cycle, or at least slows it, as long as the police costs don't rise as fast as the population.  And there's no reason they should -- there are a lot of fixed costs, and it won't take twice as many officers to patrol one square mile if the population of that square mile doubles. 

The same cycle will play out in any other city, including a suburb.  A young suburb in the early part of the cycle might seem to be in a better position -- and in a sense it is, since its day of reckoning is further away.  But it eventually will face the same reckoning.  Its police force, too, will become an ever-increasing burden, and its tax-sensitive citizens will move in increasing numbers -- unless the suburb attracts more residents over time.

Water and sewer infrastructure works almost the same way.  There is a large upfront cost, low initial expenses because young lines don't require much maintenance, but rising costs as lines age and break and water treatment plants wear out.  These culminate eventually in very large charges for digging up and replacing lines or replacing other infrastructure, which lead to a lot of hand-wringing by the mayor and the city auditor and the local newspaper.

A growing population helps with water infrastructure just like it helps with the cost of an aging police force.  A city cannot simply decree that people move within the city limits.  Most central cities don't enjoy a growing population.  But if there are people clamoring to move into the central city, the incumbent residents ought to realize that as a good opportunity to spread spread long-term costs among a larger population.

What's "The Tree" worth?

This story has generated a lot of buzz and twitter.  On Tuesday, Austin's Planning Commission denied a variance to remove a big tree on a downtown property at Bowie and W. 5th.  The denial might -- and likely will -- obstruct the construction of a tall, dense residential or office tower on a piece of property that City Council just zoned for a tall, dense residential or office  tower.  

The backup packet for the PC agenda item (TP 10617196 Bowie Street (311, 313, 315)) is here (pdf). 

I'll have more to say about this issue in later posts.  But for now, I'd like to focus on just one question:  What's The Tree worth?  We put an implict value on a tree every time we insist the tree be preserved or allow it to be removed.  Insisting that a tree be preserved means we deem the tree's benefit to the community outweighs the cost to the community of lost tax revenue and positive spillovers from the development.   And vice versa. 

It's not obvius how one should value a "protected" or "heritage" tree.  I think the valuation is intrisically subjective, which is one of the problems with the heritage tree ordinance.  But there's no way to avoid implicitly pricing the tree.   The Planning Commission implicitly valued The Tree at more than $90,000 $70,000 because it rejected the developer's offer of $90,000 $70,000 for "mitigation costs."

I'd like to hear from readers what you think The Tree is worth.  How much do you think the City should be willing to pay to buy The Tree outright?  What is the maximum amount of tax revenue the City should forego to preserve The Tree?  If the City were to sell the right to remove The Tree, what minimum price should it demand?  The three questions should yield the same answer.

I'm not interested right now whether you think the project is feasible with The Tree in place, or your estimate of the lost tax revenues, or whether you think PC has the authority to consider a tree's benefits when weighing a variance, or whether you think PC's decision complied with the heritage tree ordinance criteria.

The Tree is a pecan tree.  It is approximately 60 years old.  It is 57' tall and has a roughly symmetrical canopy with a 55' diameter spread.  It has a 32.0 inch diameter at breast height.  The City arborist rated The Tree in "good" condition, and the developer did not dispute that The Tree is healthy and in good condition.  It is set back 90' or so from Bowie Street and approximately the same distance from W. 5th St.

Here are some photos.

These photos were taken from the driveway entrance to the property.  This is the only street-level vantage from which The Tree is visible from root to crown.  I'm sure some residents of the Monarch high-rise just across Shoal Creek have a full view.

Here are a couple of closer shots:

Here is aphoto of the property looking south from the other side of W. 5th.  The small buildings lining W. 5th are not part of the property at issue and would not be torn down to build the project.

What say you?

NB.  Because valuations like this are intricially intrinsically subjective and personal, there are no right or wrong answers.

Another NB.  The Tree cannot be moved to another site because its roots are too deep, if that affects your valuation.

Sentence of the day

[T]he optimal urban area size is about 750,000 people and the optimal density is about 2,000 jobs per square mile–roughly Baltimore or Philadelphia.

That's from Randal O'Toole.

I did not know that.  The Census Bureau has not yet released "2010" census figures for urban areas, but in 2000 there were at least 33 urbanized areas with too many people. 

(The paper O'Toole cites does not, in fact, attempt to calculate the population that maximizes productivity but rather the population that maximizes patent intensity per capita.  See The Gated City for the extensive evidence that productivity rises with density.)

Ryan Avent has a detailed response to O'Toole here.

Urbanists should care about taxi cartels

They really, really should.

The main constraint on denser commercial and retail on streets like South Lamar is not traffic, but parking.  An architect friend told me that a surface parking space needs a minimum of 325 sf.  That's a lot.  It's so much that many of the small strip malls on South Lamar wouldn't be redeveloped even if parking minimums were eliminated.  (The City imposes other significant hurdles to redevelopment, of course.) 

If every trip is by car, then we need more parkings spots per car.  The right way to think about taxis is not as an alternative to a car trip, but as an alternative to parking. 

Taxis are also important to people who want to go car-free.  These people need a range of travel options.  Buses and bikes aren't enough.  Car2Go and more traditional rentals fill a niche.  But taxis are important, too.  Some trips don't warrant renting a car but are too cumbersome to make by bus.

So a healthy urban environment needs reliable taxi service at reasonable cost.  The City makes taxi service less reliable by making taxis scarcer, and it makes the rides more expensive.  It thereby encourages more parking, albeit indirectly.  Urbanists ought to care about this.  They ought to complain.

Austin's taxi cartel

The Austin City Council restricts the number of city taxi permits (to 668) and restricts who can get them (just three companies).  It thereby enforces a taxi cartel.  The City Council just received the report of its taxi consultant, Ray Mundy, whom it paid $50,000 to advise it on (1) how to strengthen the cartel, and (2) how to divvy up its spoils.  Most of the debate in town has centered on (2).  Apparently, there is no public consensus on how to divide the surplus profits generated by a cartel-- probably because in most industries, cartel members go to jail once the cartel is discovered, which makes them reluctant to conduct public discussions of cartel profits. 

The nascent pedicab industry in downtown Austin must be cutting into taxicab profits.  Or maybe not.  Perhaps Mundy simply spotted a free market in need of old-fashioned price fixing and market allocation schemes: 

As for pedicabs, Mundy said Austin has too many companies offering the service, an uncertain fare "system" and a good number of unsafe vehicles.

He suggested that the city stop issuing permits to new companies and somehow encourage that industry to consolidate.

Mundy also said that the city should allow only pedicabs specifically manufactured as a unit, banning the makeshift type in which a passenger cart is attached to a regular bicycle with some sort of bar. And the city should set a minimum fare, with tips possible, rather than the current Wild West situation where drivers basically charge whatever the market of the moment will bear.

I'm flabbergasted.  I'm flabbergasted by the brazenness of the report, and I'm flabbergasted by the public's apparent indifference.  Cartels are bad for everyone who is an industry outsider.  They are bad for would-be entrants into the market, and they are bad for consumers.  This timely post by Matt Yglesias on Milwaukee's cartel is directly on-point:

The city of Milwaukee capped the number of taxi permits at 321 way back in 1992, and consequently the price of a permit on the secondary market has risen to $150,000.  That’s excellent news for people who made well-timed investments in Milwaukee taxis, but obviously it’s bad news for would-be taxi entrepreneurs who’d like to get into the market. Fewer permits means fewer jobs for cab drivers, and it means less access to taxis in the city.  Dynamics of who uses taxis, what they cost, and where they’re available differ a lot from city to city and I’m not familiar with Milwaukee so I can’t say who in particular is disadvantaged by taxi scarcity but it’s not the sort of thing that helps a city’s quality of life or economy.

I'd add that taxi-rigging is also bad for the city's culture.  It fosters, if not outright corruption, then a constant suspicion of corruption, since it gives Council the  power to hand outsized profits to a favored few.   Not suprisingly, both taxicab companies and taxicab drivers (who believe the cartel should consist of incumbent taxicab drivers rather than incumbent taxicab companies) poured tens of thousands of dollars into the last city council elections.

I think that if it is a crime for pipefitters and vitamin makers to fix prices and allocate market shares, then it should be a crime for city councils to do the same on behalf of taxi companies.  It is not a crime, though; the market-rigging merely provides a nice fund-raising opportunity.

Eliminating the antitrust exception for government action would do wonders for municipal governance. 

NB.  Some people misinterpret me to advocate the elimination of licensing.  I do not.  Taxi drivers should be required to show the necessary state-issued licence (Class C?), provide their driving history, etc.   The City can and should establish minimum qualifications.  The City should not use a licensing regime to limit the number of drivers or their rates, though.

Commenters have suggested all sorts of hypotheticals in which dishonest taxi drivers fleece hapless riders or do other bad things.  Why this is a particular concern with taxi drivers escapes me.  In any event, the City can always pull the licenses of drivers who rip off customers or otherwise break the law. 

Randal O'Toole reviews The Gated City

Here:

Didn’t we read this book already? Only last time it was called The Triumph of the City and it was written by Edward Glaeser, a fine economist whose view of cities is slightly shaded by the fact that he grew up in Manhattan.  Avent’s bio doesn’t say whether he grew up in New York [AC: Raleigh, NC], but he did go to school in London and currently lives in Washington.  I can’t help but feel he is another city boy who loves cities so much he wants to impose them on everyone else.

The Antiplanner would never step into another city again if I didn’t have to as a part of my work.  But I don’t think everyone would be better off living my lifestyle; I just think people ought to have a choice.  Avent’s book simply stimulates the smart-growth advocates who want to impose all sorts of policies on urbanites and suburbanites to force them to live in denser communities.

I infer from this that (i) O'Toole didn't read the book; (ii) he didn't read the book carefully, or (iii) he did read the book carefully but is misrepresenting the conclusion.  Ryan does not advocate any coercive policies in The Gated City.  He asks that the housing market be allowed to function like it's supposed to:   the market should be allowed to supply more housing when and where there is demand for more housing which means -- gasp -- places should be allowed to grow denser when there is demand for it.  It's sad that it takes an economics-laden, research-intensive book to justify this principle, and even sadder that so many self-described libertarians denounce this principle, but such is the state of land-use policy in the United States today.

Like O'Toole, "I just think people ought to have a choice."  People do have different preferences.  Some people like dense, bustling, noisy places.  And some people -- including, by his own admission, O'Toole -- hate them.  The problem is that the people who hate cities try to impose their preference on the places where cities want to be. 

After all, only a small portion of even the densest metropolitan area is genuinely urban.  Most "city" dwellers actually live in places that are relatively sparesely populated.  This would still be true even if the law allowed the demand for urban places to be fully satiated, even in places like San Francisco and San Jose.  People who hate dense environments have and will always have thousands of suburbs and small towns to choose from.  If you don't like dense places, don't live in dense or densifying places. 

The libertarian position accordingly ought to be that we should allow the market to provide density where and when people want it; the people who don't like that particular product can buy a different product.  That occasionally might require a density-hating libertarian to accept change to his own neighborhood.  It occasionally might require him to move to another neighborhood if he finds the change too uncongenial.  But libertarians are supposed to embrace dynamic markets.  They deride people who fight market-driven change as "statists."   Don't preach creative destruction for me but not for thee.

The view that we should accommodate density ought to be the progressive position, too.  Allowing housing to keep up with demand keeps home prices low, which certainly is more equitable and egalitarian than propping them up.  Allowing housing to keep up with demand lets people move freely from low-opportunity cities to high-opportunity cities.  Allowing housing to keep up with demand facilitates the density necessary to support transit, preserves open space on the city's periphery, and probably reduces VMT and the other things that environmentalists should care about.  Fighting to preserve a neighborhood just because you happen to live there and like it the way it is is not only statist, it's selfish, even if you can't identify the specific people who will be hurt.

I've always thought that the libertarian position and the progressive position on land-use policy ought to be approximately the same.  Thus it is not surprising per se that so many "libertarians" and "progressives" agree on land-use policy.  What is strange is that so many "libertarians" and "progressives" agree because the libertarians turn into command-and-control statists  and the progressives into regressive elitists.  

Land prices vs. entitlement prices

Matt Yglesias quibbles with The Gated City because "it’s a perpetuation of the American habit of mixing up discussions of the price of houses with discussion of the price of land":

Land is a speculative commodity, comparable to a bond or a stock or oil futures. You might lose money buying Texas Hill Country land and you might make money. The house is more like a bulldozer or a camel. It’s expensive, and its potential resale value is a relevant issue when you’re considering a purchase, but it’s not an investment. Five years from now, your house, like your bulldozer, is going to be older, more broken, and cheaper than it is today. It’s the land the house/bulldozer is on that might be more expensive. That might be because it becomes more desirable as a place to live, or because climactic shifts make the lend better-suited for high-value agriculture, or because oil is found on the property, or because it acquires a better road connection, whatever. There are lots of perfectly ordinary reasons for land to go up or down in price. A house, by contrast, is a large decaying physical object.

This is a reasonable point but I don't know what it has to do with Ryan's thesis.   The price of a piece of land actually has three components:  (1) the value of the physical fixtures, (2) the price of the land, and (3)  the value of the things you can legally do with the land.

You might think that (2) and (3) are the same but they are not.  Suppose we calculated the value of all land in Los Angeles by subtracting the value of all structures and improvements from the market value of the land.  The number we got, of course, partly would reflect what the city lets people do with the land.  Now suppose we calculated what the value would be if the City let people build as densely as they wanted.  Those numbers would not be the same.  The difference would be due to restrictions on development.  (The first number would not necessarily be higher than the second for the same reason that the total value of the output of a monopolized market might be smaller than the total value of the output of a perfectly competitive market.) 

This raw calculation is intractable so let's take a more tractable example.   A 47-story condo tower uses just as much land as a 46-story condo tower.  If you could build as tall as you wanted, and held the price of land constant, then the price of the condos on the 47th floor should equal their cost to build.  But Glaeser and Gyourko showed that in Manhattan in 2004, the price of the marginal condo fetched as much as a $300 psf premium over the cost of construction.  That premium reflected the value of the entitlement, not the value of the land.   

Obviously, the value of the right to build can rise or fall with demand, just as the value of land can rise or fall with demand.  So why draw a distinction between the value of land and the value of entitlements?  Because the supply of land is fixed; the supply of entitlements is not.  In an efficient, equitable world, a city would set the price of the entitlement to build one housing unit equal to the social cost of that housing unit.  In high-value, high demand places, where additional people produce net positive returns, the value of the entitlement should be zero or close to it.  Instead, it is worth hundreds of thousands of dollars.  That inflated price produces dead-weight losses for the residents of the city and the country as a whole, and that's what Ryan, Glaeser and a host of others are complaining about. 

I'm pretty sure Matt understands this, but his post unfortunately will lead to a perpetuation of the American habit of mixing up discussions of the price of land with the price of entitlements.