Ninety minutes of highway separate Tyler, Texas from Sherman-Denison. Both sit inside the Texas Triangle. Both register demand running ahead of price, and a screen would surface them together.

Then you count the houses.

Sherman-Denison has 147,000 people and permits about 1,500 homes a year. Tyler has 245,000 — two-thirds larger — and permits about 580. The smaller market is building two and a half times the housing.

Sherman-Denison sits north of Dallas in the corridor absorbing the semiconductor investment that has landed there, and its numbers behave accordingly: population growing near 2.4 percent, employment near 1.7, permits above ten per thousand residents — a higher rate than Dallas-Fort Worth’s. Demand is arriving and the market is answering it.

Tyler is growing too. Population up better than 1.5 percent, employment up 1.2, in-migration positive. It is the cheapest of these markets at about $354 a square foot, and its renters spend under 29 percent of income on rent, which is comfortable. On every measure of demand and affordability Tyler looks like a market that should be building. It produces 580 homes a year.

So what is happening in Tyler? This dataset cannot tell you, and it is worth being precise about what it rules out. Not a demand problem — people are moving there. Not an affordability problem — the price and burden figures are the healthiest in this group. Not an employment problem. Whatever is throttling production sits in the machinery: lot supply and who controls it, a wastewater or utility constraint, an approval calendar, how much development sits inside the city limits versus out in Smith County, or the absence of a builder large enough to run a pipeline. East Texas land ownership is old and concentrated, which is a candidate and not a finding.

A builder arriving in Tyler with capital and a land team would find the demand exactly as advertised and the production impossible, and would spend a year and real money learning which explanation was the operative one.

This is the specific way regional allocation fails a homebuilder. The builder’s exit is a household with a mortgage, and that exit requires finished lots. Lots are manufactured one jurisdiction at a time, by people who answer to voters in that jurisdiction and to nobody else. No regional variable predicts a planning commission.

Waco is the useful inversion. Among the four Texas Triangle markets here — Tyler, Sherman-Denison, Dallas-Fort Worth and Waco — Waco has the strongest employment growth at 1.7 percent, ahead of Dallas-Fort Worth’s 1.3 and Tyler’s 1.2. It permits at a healthy 5.8 per thousand. And almost nobody is moving there: net domestic migration near zero, population growth under one percent. Waco has built the capacity and is waiting for the households.

That is a better entry position than Tyler’s, and the reason is time. Idle production capacity can be bought cheaply and switched on when demand arrives. Blocked production stays blocked until the politics change, and politics do not run on an investment horizon. One is a timing problem. The other is structural.

The majors sit where you would expect — Dallas-Fort Worth and Austin both producing heavily into real population and payroll growth, and priced for it, Austin near $503 a square foot against Tyler’s $354.

So the order of operations is fixed, and it is not the one most allocation decks use. A regional read narrows 402 markets to a shortlist. It cannot rank that shortlist. The fee schedule, the approval calendar, the utility district’s capacity, who sits on the council and what they campaigned on — all of it sits below the resolution of any regional dataset, and has to be established market by market, before land is tied up rather than after.

A builder who allocates to “the Texas Triangle” has named a price regime and answered nothing.

Demand tells you where to look. Entitlement tells you where you can go. Never let the first stand in for the second.

Each quarter we score all 402 U.S. metros on two layers — a production layer built from permits, migration and employment, and a price layer built from how prices and rents have behaved — each as a percentile against every other market. The gap between them is the reading. Megaregion names follow Yaro, Zhang and Steiner (Lincoln Institute, 2022); metro assignment is our own reading of their published map. Tyler, Dallas-Fort Worth and Austin are high-coverage at a 2025-Q4 vintage; Sherman-Denison and Waco are partial coverage at 2025-Q4 and directional only.

In April 2025 this blog closed a piece with “Boise is still booming, and the builders are right in the middle of it all.” With the 2026 data in, booming is the wrong word. Boise is holding a balance — and someone has to keep holding it.

CESE, this desk’s market-signal engine, reads 402 metros on two layers — forward, built from permits, migration and employment; confirmation, built from prices and rents. Forward far ahead of confirmation reads early — production signals promising more than prices admit. Far behind reads late — prices coasting on momentum the indicators no longer support. Boise, on the 2026-Q2 release, scores 70.6 forward against 65.2 confirmed — five points apart — barely a gap. Both layers strong, neither ahead. That is NEUTRAL — and the drivers say who has to hold it.

Nothing about Boise’s price says anything is happening. Ada County closed 1,001 single-family sales this May against 845 a year earlier — 18.5 percent more volume — and the median slipped seven-tenths of a percent, to $575,900. Days on market barely moved. What moved is the shelf: inventory down 13.6 percent, months of supply from 3.0 to 2.3. Boise cleared a bigger year through a smaller store without touching the price tag. Eighteen percent more buyers absorbed at a flat price means the supply side is producing enough to meet them. The builders are keeping the price flat — and only while they keep pace.

Two forces are loading against that stride. The first is in southeast Boise, where Micron is building two fabs — $50 billion, 3,500 direct jobs, roughly 17,000 statewide with partners, by its own numbers. The build is already in the data: construction payrolls grew 4.8 percent in the year through June, four times the metro’s overall rate. The hiring is not: manufacturing employment grew 0.3 percent. The first fab finishes in 2026 and starts producing in 2027. The buyers are scheduled. They haven’t closed.

The second force is City Hall, which has spent two years dismantling its own gate. A rewritten zoning code took effect in December 2023, moving infill and ADUs to approval by right instead of by hearing. ADU applications went from just over three a month to twenty-seven. The council loosened again this March, unanimously, and is master-planning around the fabs. “Boise is not full,” the city’s long-range planner says. But an open gate is not traffic: 1,962 units approved in 2025, against 3,373 in 2022. The drop is not the city’s doing. The 2022 approvals were underwritten in cheap money; then the Fed took rates from near zero to above five percent, and deals that penciled in 2021 money stopped penciling. A developer who paused a project near downtown in early 2024 said it plainly: “construction costs and interest rate escalation” had “caused multiple project stalls across the Treasure Valley.” His zoning kept moving. The money didn’t. The 2025 count is the recovery, not the decline — up from 1,802 in 2024. The city removed its half of the friction; capital has been slower coming back with its half.

On the demand side: 18 percent more buyers, with the fab hiring still ahead of them. On the supply side: a thinner shelf every month, approvals under peak, and 682 multifamily permits that must become deliveries before badge-ins begin. If approvals hold near two thousand while the fab staffs, the shelf empties into the payroll wave — prices stop being flat, and whoever tied up dirt at a 2024 basis collects the spread. If the by-right permits convert to rooftops in time, the wave lands on supply, prices stay pinned, and the spread goes to whoever executes. A slipped fab date flips the board: construction payroll rolls off before replacement jobs arrive, and 2.3 months of supply finds out what loose feels like.

The evidence leans to the first. The shelf is thinning before a single production hire, and a permit takes longer to become a house than an offer letter takes to become a buyer. Boise’s equilibrium is real, and temporary — the balance a market holds while both sides load. My read: the payrolls land first, and the sellers win the break.

Sources: Boise Regional REALTORS, Ada County SFH May 2026 stat sheet (data as of 2026-06-19); BLS, Boise City MSA Economy at a Glance, June 2026; Micron statements as reported by KIVI-TV, 2026-02-25; BoiseDev 2026-02-18 and 2026-03-11; BoiseDev 2024-02-13 (developer quote); Next City 2024-04-25; C&C, CITY TO WATCH: Boise, 2025-04-28; CESE 2026-Q2 release (vintage 2025-Q4).

Irving’s answer to the State of Texas was eighty-five feet. Under the city’s August 2025 ordinance, apartments must stand at least that tall, with eight occupied stories. This was not defiance of Senate Bill 840, the state law that had just taken the city’s power to refuse them. It was compliance, with dimensions. Irving did not refuse. It specified.

The law it was answering passed in the 2025 session and took effect September 1. Its premise is arithmetic: the big Texas metros keep adding people, the zoning maps of their suburbs were drawn for office parks and retail strips, and much of that commercial land now sits underused while apartments wait on rezonings. SB 840’s answer was to stop making housing ask. In the covered cities — over 150,000 people, in counties over 300,000 — housing became a use the commercial map already permits.

What the bill does

• Multifamily and mixed-use residential (at least 65 percent residential by floor area) are allowed by right wherever office, commercial, retail or warehouse use is allowed. No rezoning, variance or special approval.

• Cities cannot cap density below the greater of their own highest allowed density or 36 units an acre; height below the greater of what the commercial zone allows or 45 feet; or setbacks tighter than the lesser of the commercial standard or 25 feet. No FAR limits.

• Local judgment survives where it should: heavy industrial zones and land within 1,000 feet of them, land within 3,000 feet of an airport or military base, and clear zones are carved out.

• Office, retail and warehouse buildings at least five years old, converting at least 65 percent to residential, cannot be charged traffic-mitigation fees or new impact fees.

• Enforcement lives in the statute: any aggrieved person or housing organization can sue a non-complying city, and a prevailing claimant is awarded costs and attorney’s fees.

What the cities wrote back

Cities answered in ordinance language — minimums where the state wrote maximums. Two examples:

• Irving: every multifamily building at least 85 feet and eight occupied stories; mixed-use projects at least 35 percent non-residential; 20 percent of the site as open space; in projects over 25 units, no more than half the units with the same bedroom count.

• Frisco: multifamily at a minimum of 15 stories and 150 feet unless the site sits more than 200 feet from single-family zoning; ground floors built to retail standards.

What it built

The first count arrived July 29, from Texas Appleseed, the American Enterprise Institute and Texans for Reasonable Solutions: 8,764 units in the law’s first nine months, 1,529 of them in projects filed directly under the bill’s provisions. The authors call the count a floor, not a full measure. The distribution is stranger than the total. Plano — a city the report sorts with the resisters — produced 827 direct units across three projects, one of them 475 units. Irving produced none. Two cities took the same posture toward the state, and only one of them left dimensions a building could finance. The posture is not the variable. The dimensions are.

The instructive part is that both sides still hold real cards. The state kept the industrial buffers and flight-path setbacks in the bill itself — it took the veto, not the judgment. The cities kept the form of the building, and some are using it to zone by another name. The line between those two gets tested parcel by parcel, by whoever has money in the ground, because the enforcement right sits in the statute rather than in Austin. California runs the opposite machine: the remedy there is the attorney general, who sued five cities over their housing elements on July 16. Texas’s version is slower-looking and healthier — an argument conducted in ordinances and filings, between parties who each have to defend a specific number.

The next move is already drafted. At the July 21 House hearing on these ordinances, a Texas Public Policy Foundation scholar proposed amending Chapter 2007 of the Government Code — which today exempts municipal action almost entirely — so that a downzoning that cuts property value triggers a claim, on the Arizona Proposition 207 model: repeal the rule, issue a binding waiver on the parcel, or pay. My read is that it would end the argument rather than sharpen it. A compensation claim prices a height minimum; it does not calibrate one. A city that must price every dimension it writes will stop writing dimensions, and the local half of this conversation goes quiet — including the parts the legislature itself chose to keep.

A preemption regime is worth two things: what it leaves the city to decide, and who besides the state can test the line. Texas wrote both into the statute and is now arguing over the settings. That is not dysfunction. It is the conversation every state and its cities should be having — and most are not having at all.

Sources: Texas Local Government Code ch. 218 (read at source); City of Irving ORD 2025-11165; Winstead client briefing 2026-03-05 (Frisco detail); Texas Appleseed / AEI / Texans for Reasonable Solutions report, 2026-07-29; The Texan, House Land & Resource Management hearing, 2026-07-21; Texas Government Code §2007.003(b)(1); Ariz. Rev. Stat. §12-1134; California AG release, 2026-07-16.

Armando Carbonell, who gave the phenomenon its name, has been careful about what a megaregion is. He calls it a construct and a tool — a way of thinking about space rather than a fixed thing sitting out on the ground, and something that changes as the country does.

That is the right frame, and it invites the obvious question. If a megaregion is a tool, what is it good at measuring, and what does it miss?

The current standard is the thirteen-megaregion map in Megaregions and America’s Future, published by the Lincoln Institute in 2022: the Northeast, Piedmont Atlantic, Florida, Great Lakes, Gulf Coast, Central Plains, Texas Triangle, Front Range, Basin and Range, Cascadia, Northern California, Southern California, and Arizona’s Sun Corridor. Its authors note that better than eight in ten Americans live inside these boundaries, and that they account for more than ninety percent of the national economy. That is not a niche framework. It is how a great many allocators now describe the country.

We ran our signal work across all 402 metro areas and sorted them into the thirteen. Every market reads on two layers, and the distinction is the whole argument. The production layer describes whether housing is being made and whether anyone is arriving to buy it — permits per thousand residents, net domestic migration, employment growth. It is the builder’s world. The price layer describes what has already happened to values, how prices and rents have behaved against every other metro in the country. It is the lender’s world, and the appraiser’s.

Take Detroit and Columbus. Both are Great Lakes markets, both are large Midwestern metros, and on the price layer they land within two points of each other. Underneath they are not the same economy. Detroit’s employment base is contracting. It is losing residents. It permits under two homes per thousand people — roughly 8,600 units a year for a metro of 4.3 million. Columbus, half Detroit’s size, permits nearly six per thousand and about 12,900 units a year. Half the population, fifty percent more housing. It is gaining residents and adding jobs. One of these places has wound its housing production down; the other is running it. The Great Lakes boundary contains both and distinguishes neither.

Boston and Harrisburg-Carlisle are the harder case, and both are Northeast markets — Harrisburg sits at the western edge of the corridor where I-81 crosses I-78, three hours from Manhattan. They belong to the same megaregion by design: the map is drawn on economic linkage, and Harrisburg’s warehouses exist to serve the Northeast’s consumers. That is exactly why the comparison is fair. These two are supposed to belong together.

They permit at almost the same rate — Boston 2.126 per thousand, Harrisburg 2.294. Everything else diverges. Boston is a knowledge economy: average wages near $90,000, household income near $116,000, property around $500 a square foot. Harrisburg is a distribution economy carrying more than four times Boston’s logistics employment per capita, paying about two-thirds the wage, at $311 a square foot.

Push on why neither builds, because the reasons are not the same. Boston loses 30,000 domestic migrants a year and still grows, meaning its demand arrives from abroad and from natural increase, onto the most expensive land in the region. Harrisburg gains domestic migrants, grows population faster than Boston, sits on cheap land, and still produces under 1,400 units a year. Boston looks like a market where demand exceeds what approvals and land cost permit. Harrisburg looks like a market where demand exists and the production apparatus is simply small — few large builders, a modest lot pipeline, a county-scale development industry.

That distinction is inference, not measurement. The dataset holds no approval calendars and no lot supply. Which is the point: the thing separating these two markets is invisible at the regional level, and it is the only thing a builder needs to know.

Across the 370 markets we can signal on, the pattern is consistent. Grouping by megaregion cuts the variance of the price layer by 44 percent against random groupings of the same sizes. It cuts the production layer by 17 percent. Both results are strong, and the difference between them is the finding: the boundary is about two and a half times better at enclosing a price regime than a production regime. Tell me a market’s megaregion and I can guess its price position reasonably well, and its construction rate barely at all.

There is a mechanism behind this. Price travels and entitlement does not. A comparable sale in Fort Worth informs an appraisal in Dallas. A lender running a Sun Belt allocation prices Tampa and Jacksonville off one curve. Capital markets, labor markets and appraisal practice arbitrage value across linked metros — that arbitrage is most of what economic linkage means when you look closely. Entitlement arbitrages nowhere. A rezoning in Frisco does nothing for Denton. Impact fees, approval calendars, moratoria, sewer capacity, the composition of a planning commission: set one jurisdiction at a time, stopping at the city line.

One objection, and it is fair. These boundaries were drawn partly on density and economic integration, and neither is independent of price. Some of the coherence inside them may be an artifact of how they were drawn. That cannot be settled from one quarter of data. It weakens the explanation and leaves the consequence intact.

Carbonell is right that the megaregion is a tool. It is a precise one, and it measures value.

The megaregion is a price regime. It tells you what an exit will be worth. It tells you nothing about whether you can build anything to sell.

Each quarter we score all 402 U.S. metros on two layers — a production layer built from permits, migration and employment, and a price layer built from how prices and rents have behaved — each as a percentile against every other market. The gap between them is the reading. Megaregion names follow Yaro, Zhang and Steiner (Lincoln Institute, 2022); the assignment of individual metros to those regions is our own reading of their published map, not their county-level definition. Markets cited here are high-coverage at a 2025-Q4 vintage. Components: Census BPS, Census PEP, BLS, ACS 5-Year, FRED, each at its own vintage.

When Lennar acquired WCH Homes in 2024, the principal did not take the check and leave the industry. Dylan Wingate sold the homebuilding operation and formed Patriot Development Group to supply finished lots to Lennar and every other large builder in South Atlanta. His stated reasoning was that land development was where he could generate the most value going forward.

That is a man who read the consolidation wave correctly and moved one step upstream on purpose.

Why the builders are merging

KB Home’s gross margin fell from 21.2% to 19.7% between the second quarters of 2024 and 2025. Lennar guided Q1 2025 margins to 19.0–19.25%. Dream Finders’ SG&A rose 39% year over year to $135 million.

Scale is the response. A larger builder spreads fixed SG&A across more closings and buys trade labor at better terms. This is scale as a margin defense, not a margin driver. Nobody expects the combined entity to out-earn what the acquirer earned three years ago. The alternative is watching margin compress with a smaller base to absorb it.

The buyers winning these auctions have the least urgency. Sekisui House paid $4.9 billion for M.D.C. Sumitomo Forestry paid $47.00 per share for Tri Pointe — a 29% premium for a company entering the deal with 155 communities and a healthy backlog. Berkshire agreed to Taylor Morrison at $72.50 cash. None of them needed the deal accretive next quarter. The constraint on domestic acquirers here is not regulatory. It is cost of capital and return timeline.

What they bought was territory: Toll Brothers took nine communities and 1,500 lots in Northwest Arkansas. Daiwa House assembled a Southeast platform one regional builder at a time. Lennar bought Rausch Coleman for affordable volume across five states. The lot pipeline is the asset.

The part worth paying attention to

Taylor Morrison closed a $3 billion Kennedy Lewis land and construction facility in July 2025 — months before Berkshire’s offer. The land-capital problem was solved by a third party before the company was ever acquired.

Dream Finders went further. Its Beazer proposal arrived with Millrose Properties already committed to acquire Beazer’s homesites at closing. Dream Finders is not proposing to buy Beazer’s land. It is proposing to buy the operating business and have a REIT hold the dirt.

Millrose exists because Lennar built it — $5.5 billion in land spun into a REIT in February 2025, now serving Lennar and eleven other builders across 250,000 homesites. A vehicle one builder created to lighten its own balance sheet is now acquisition financing for a competitor’s hostile bid.

What that does to the land developer

The obvious read — more scaled counterparties, tighter terms, worse outcomes for whoever develops the lots — is wrong.

Land banking represents roughly $87 billion in annual finished lot value. Lot bankers have deployed over $40 billion in equity at unlevered IRRs of 10–14% on short-duration positions, 18–25% on longer-duration raw land. That capital does not replace the land developer. It backs the developer, who still holds a piece of the deal. A developer funding horizontal work with expensive regional bank debt and personal guarantees three years ago, now accessing an institutional facility, did not lose leverage. His return on the same dirt went up.

The two events are not opposing forces. Builders are shedding land because public markets pay for asset-light structures. That land has to go somewhere. It is going to developers being handed cheaper capital to hold it — and more of the risk, since sequencing and delivery now sit with them rather than a vertically integrated builder. The lot purchase agreement remains an option on the builder’s side and a forward commitment on the developer’s. Consolidation did not change that asymmetry. It changed who holds the option, and the new holders have more capital duration and less attachment to any single submarket’s schedule.

None of which fixes a bad basis. A developer who bought correctly still has margin. One who chased a takedown schedule into an aggressive land price because a builder waved a big name at him does not.

The acquirers are paying premiums for lot pipelines they intend to hand to a REIT the day the deal closes. That tells you which end of the production chain holds the scarce asset.

SpaceX priced at $135 on June 11 and closed its first session at $161.11, valuing the company above $2 trillion. The offering minted thousands of employee millionaires company-wide, but the count concentrates where the headcount does. Analysts covering Brownsville expect up to 4,000 new millionaires near Starbase. Austin’s share is smaller: the Bastrop campus employs roughly 1,600 people, heading past 2,000, and production-heavy sites carry thinner equity grants than engineering hubs. Long-tenured staff have also been selling into secondaries for years, converting wealth before the IPO could mint it. Scaling from headcount and grant structure, a defensible estimate for the Austin orbit is several hundred new millionaires — call it 300 to 800. That is an inference, not a count.

Then the geography splits. Ask where the money points and the answer is west. Agents report post-IPO demand for Lake Austin and Lake Travis condos, Westlake, Bee Cave, and Hill Country ranches. The western prime corridor — Westlake, Tarrytown, Barton Creek — was leading Austin’s luxury recovery before the listing. And the west has room: Austin’s $2 million-plus segment carries roughly sixteen months of inventory and 111 days on market. IPO wealth arriving into that overhang is absorption, not displacement. Nobody gets priced out of Westlake. Westlake was built for exactly this buyer, and it has been waiting three years for him.

Ask where the jobs point and the answer is east. The Bastrop campus is mid-expansion: a $280 million buildout backed by a $17.3 million state semiconductor grant, a permitted 1.1 million-square-foot solar cell factory, and an announced 11 million-square-foot satellite plant that would out-size the Tesla Gigafactory down the road. A $1.4 billion data center is rising in Cedar Creek. Bastrop County’s population grew from about 98,000 in 2020 to roughly 115,000 in 2024, and the school district is building two elementary schools to keep pace. Meanwhile the county’s median sale price is $333,000, down 5.3% year over year, on 137 days of market time.

So the eastern conversion is real, but it is not millionaire-driven. It is payroll-driven, and it predates the IPO. What the listing changes east of SH-130 is capitalization and signal: the $75 billion raise funds the factories, and land investors position ahead of the hiring. The displacement risk to renters and manufactured-home households on unincorporated county land comes from that assembly — builders and land funds buying ahead of employment growth — not from engineers buying ranchettes. Migration data supports the demand read: more Bastrop buyers currently search to leave the county than outsiders search to enter it. People move there for the work, not the place. If assembly accelerates, the outflow runs one price tier down the same highways — Smithville, Paige, Dale, Red Rock — though that remains inference; no data yet captures it.

The corridor will develop. Wealth buys views; payroll buys proximity. The prediction: the IPO shows up in west Austin closing prices by early 2027, and east of SH-130 only as land assembly and rooftops chasing shift schedules. Nothing out there yet converts either force into a town — and until a main street exists that someone would stay for after quitting the job, the east side will keep building housing for a workforce and appreciation for speculators. Austin appears to be priming for its next boom.

For most of the twentieth century, the federal tax code preferred old buildings gone. Depreciation rewarded new construction; keeping and repairing an old structure earned a thinner deduction and a maintenance bill. Run the numbers honestly and the warehouse, the depot, the mill were worth more as cleared sites than as standing assets. The code was not neutral. It paid for the bulldozer.

You can see the result in what came down. Penn Station was demolished beginning in 1963 — a 1910 landmark torn out to monetize the air rights. It was legal, rational, and by the financial logic of the moment, correct. That is the whole problem in one transaction.

The fix was not sentiment. It was math.

The National Historic Preservation Act of 1966 built the regulatory scaffolding — the National Register, the state preservation offices, the federal review. But regulation can only slow demolition. A building protected from the wrecking ball but unprofitable to repair is a slower kind of loss. The turn came in 1976, when Congress first put rehabilitation incentives in the tax code, and in 1981, when a 25% credit usable against ordinary income pulled real capital into old buildings. Investment climbed from under $1 billion to $2.4 billion in four years.

Then 1986 cut the credit to 20% and barred it against ordinary income. Demand collapsed from three thousand projects a year to a few hundred. The lesson holds: the credit’s value was never the percentage. It was who could use it, against what income, and how fast.

That 20% credit still runs today, alongside roughly 38 state programs that stack on top of it. Through 2024 it has helped rehabilitate more than 50,000 buildings — most of them not trophies but the corner blocks, schools, and depots that give a downtown its edges, kept standing and put back to work. Roughly half sit in low- and moderate-income neighborhoods. The credit does what aged buildings do: it preserves the cheap, adaptable space that lets a mix of incomes survive in the same few blocks.

The mill states wrote the sharpest version. When textiles left the Carolinas, they left hundreds of empty buildings behind. North Carolina’s mill credit turned Gastonia’s vacant Loray Mill into several hundred apartments. South Carolina went further — its textile credit lets demolition, remediation, and new construction all qualify. That breadth makes contaminated mill sites financeable. It also turns a preservation incentive into a development subsidy: the mill can be saved or torn down, and the money flows either way.

Set the design questions aside, though, and the deeper point holds for every version of the credit: this is the optimal way to redevelop. The infrastructure is already in the ground — streets, sewers, rail, the grid — and the structure is already standing. Rehabilitation reuses all of it, which is why a dollar spent here buys more local jobs and skilled trades than a dollar of new construction, and why the program returns more to the Treasury than it costs. It puts housing where the bones were built for it, and it can pay for a contaminated site’s own cleanup along the way.

The case for it was never preservation — it is arithmetic: more housing, more jobs, and a restored tax base, built on infrastructure already in the ground, for less public money than the same square footage on bare land. Redevelopment rarely gets to be both the cheap option and the right one; here it is both.

A land deal that once began with someone walking a parcel now begins with a query. The builder pulls up lot-level data — what is selling nearby, at what price, with what incentives, at what absorption rate — and the screen decides whether the dirt pencils. On May 29, CoStar agreed to pay $800 million in cash for the company that owns the screen.

That company is Zonda: more than 3,000 customers, 104% net revenue retention, and a proprietary, lot-level database used, in CoStar’s own words, for underwriting, land strategy, capital allocation, and forecasting. That is not a marketing tool. It is the layer where money decides what gets built.

This is the part of financialization worth naming. It is not more capital entering housing — capital has always been here. It is the standardization of judgment. When the underwriting data consolidates into one proprietary product, the builder, the land banker, and the construction lender look through the same instrument. Capital flows toward what the instrument scores. What the dataset cannot measure — the street, the corner, the block that adapts in thirty years — goes dark, because it never enters the query.

Where this gets sharp is title and information law. Roughly a dozen states are non-disclosure: Texas, Utah, Idaho, Montana, and others where the sale price never enters the public record. Only the parties, the MLS, and private vendors hold it. So the proprietary database is worth most exactly where the public withholds the data — and that cuts two ways.

The minus is textbook. Opacity is information asymmetry. It raises the cost of price discovery, advantages whoever can privately assemble comps, and hands the perception layer to a single vendor. A state that keeps prices secret does not protect a free market. It auctions the information rent to whoever can afford to buy Zonda.

The plus is the friction itself. Where data is not universally legible, capital cannot fully converge on one signal. Local knowledge keeps its value. The Texas appraiser who actually knows the submarket still beats the model. Opacity is inefficient, but it is also the last thing standing between a heterogeneous market and a single national underwriting lens.

So the real question is not opaque versus efficient. It is who captures the information rent — a public record open to everyone, or a private monopoly open to subscribers. CoStar’s history answers which it prefers: the FTC blocked its RentPath deal, a competitor went bankrupt after litigation with it, and three antitrust class actions landed this spring, all alleging the same conduct — keeping customers’ own data from leaving.

The next concentration in real estate will not be in who owns the buildings. It will be in who owns the dataset everyone uses to decide which buildings to fund. Opacity will not save us from that. It will only decide whether the rent goes to the county or the vendor.

On Sunday a company that spent a year insisting it could find nothing worth buying paid a 24% premium, in cash, for a homebuilder trading under nine times earnings — into the most frozen resale market since 2008. Berkshire is acquiring Taylor Morrison for $72.50 a share, about $8.5 billion in enterprise value (verified; May 31, 2026). The question is what $8.5 billion bought. Not houses.

It bought a land pipeline, a financing machine, and duration.

The pipeline, from the filings: at the quarter closest to execution, Taylor Morrison controlled 75,626 homesites — roughly 6.2 years of production — of which 51% was controlled off balance sheet, through options rather than ownership (verified; Q1 2026, March 31). The FY2025 10-K shows 78,835 homesites, 54% controlled at year-end. Owned, on-balance-sheet land is closer to 37,000 lots, about 3.0 years (inference from the reported percentages). That land-light structure is the point: control the pipeline, defer the capital. A permanent balance sheet can own that optioned half outright, cheaper, than a public builder managing quarterly leverage ever could. The optioned half is the part the deal makes more valuable.

The financing flywheel matters as much: Taylor Morrison originates its own mortgages and sells its own title and insurance (verified). The home is customer acquisition for a finance annuity.

Honest flag: the pipeline has been shrinking — 86,266 lots in Q1 2025, 75,626 by Q1 2026, the optioned share falling from 59% to 51% (verified). Management was pulling back on land in a high-rate market. Berkshire is buying at contraction, not full extension — value at the trough, not a land grab at the top.

Six weeks earlier, Toll Brothers bought Buffington Homes of Arkansas — nine communities, 1,500-plus lots, its 16th builder acquisition since 1995 (verified). Same trade, smaller scale: lots and a captive operating team, financed by the stronger balance sheet.

What is being valued is now unambiguous. Land pipeline over buildings. Captive financing over construction margin. Permanent capital over public-market liquidity. The public market priced Taylor Morrison at nine times earnings; Berkshire paid a quarter more and took it private. That gap is the story.

For scale, the whole public group is cheap. As of late May 2026, D.R. Horton traded around 12–13× trailing earnings, Lennar near 12×, Toll Brothers near 10×, PulteGroup around 9×, and KB Home, Meritage, Tri Pointe, and Century Communities at roughly 5–9× — a sector range that sits far below the broader market, after more than a year of de-rated, range-bound builder stocks (verified; figures vary by source, share class, and trailing-versus-forward basis). Taylor Morrison’s sub-9× sat at the cheap end of the large-caps. The one premium is the tell: NVR trades near 16×, and NVR is the purest land-light operator — it options nearly all its lots and owns almost no dirt. The public market already pays its richest homebuilding multiple for the most capital-light pipeline. Berkshire just bought a cheaper version of that model outright and removed it from the tape.

The macro frames it. Existing-home sales are stuck near four million annualized — a 2008 level — because owners holding 4% mortgages won’t trade into 6.5% ones (verified). In a frozen resale market, the builder channel is the one still transacting. So permanent capital is buying the production system at a discount the public market refuses to close, taking the cash flows private, and waiting.

The signal for capital markets: with rates higher-for-longer and a record cash pile earning T-bill yields, the marginal buyer of cyclical assets is no longer the public shareholder. It is the balance sheet that can wait. Expect more builders — especially land-light platforms with captive mortgage operations — to leave the public market the same way: bought cheap, taken private, held for a cycle the public tape won’t underwrite.

Ezra Klein Grills CA’s Possible Governors | The Ezra Klein Show

The diagnosis is correct. California’s housing shortage is a government-made problem. Decades of local veto power, environmental review abuse, restrictive zoning, and a fiscal structure that rewards commercial tenants over residents produced a state where building anything is slower, riskier, and more expensive than almost anywhere else in the country. The candidates know this. Their disagreement is over which lever to pull first.

The Three Positions

Steyer’s answer is financial. Cities resist housing because apartments go to school and use health care — used car lots don’t. His solution: “You do it, you get the money. You don’t do it, you don’t get the money.” He pairs fiscal incentives with a construction proposal — the state purchasing prefabricated modular components, similar to an auto supply chain, and providing them to developers — to attack costs directly.

Mahan’s answer is authority. When cities don’t deliver, “the state should override and create by-right mechanisms for developers to move ahead projects whether or not the city likes it.” He points to San Jose as proof: simplified approvals, exemptions for zoning-compliant projects. He wants that model at statewide scale.

Porter’s answer is process. A single statewide development permit to streamline approvals, built on the logic that every time you have uncertainty about whether something is going to happen, the costs go up.

What Each Gets Right

Steyer correctly identifies that the fiscal structure of local government is the root problem. Service obligations outpace tax revenue from residential development, so cities block housing to protect their balance sheets. Changing that incentive — paying cities to build rather than penalizing them for blocking — attacks the structural cause rather than the symptom.

Mahan correctly identifies that local veto power has been captured by incumbents who block all density regardless of merit. His override mechanism doesn’t ask cities to want housing. It makes their opposition irrelevant. That is politically blunt and structurally honest.

Porter correctly identifies that timeline uncertainty is a cost driver as real as labor or materials. A project that takes three years to entitle carries financing costs, holding costs, and escalation risk that a six-month entitlement does not. Her permit reform attacks that directly.

The Real Comparison

The three positions reflect three different theories of where government friction is worst.

Steyer thinks the problem is incentive. Fix what cities are rewarded for and behavior follows. Mahan thinks the problem is authority. Cities will never want housing, so stop asking and start overriding. Porter thinks the problem is process. Compress the timeline and private capital closes the gap.

All three are probably right. They are also not mutually exclusive. The more useful question is sequencing: which lever produces housing fastest given California’s political structure, and which creates durable conditions for continued production rather than a single-cycle spike.

Steyer’s incentive model is slower but more durable. Cities that build because it pays will keep building. Mahan’s override model is faster but depends on sustained political will at the state level to enforce it. Porter’s process reform is the most market-neutral — it doesn’t pick winners or override local governments, it just removes friction and lets capital decide.

California built its way into a shortage through fifty years of government obstruction. The next governor’s job is to dismantle that obstruction and step back.

The market will determine what gets built. Government created this problem by refusing to get out of the way. The solution is not a new set of government preferences. It is the removal of the old ones.

In most of America, a developer who wants to build a thousand homes on raw land faces an infrastructure problem with no clean solution. The water lines, sewer mains, drainage channels, and roads must exist before anyone moves in, and none of it generates revenue until it does. Cities will not fund it — they answer to existing voters who see no reason to finance infrastructure for people who have not arrived. Counties lack the bonding capacity. State and federal programs are too slow.

Texas looked at this problem and, starting in the 1960s, built a layered system of special-purpose districts that allows infrastructure to be financed, built, and paid for by the people who benefit from it, without requiring cities or counties to carry the risk. Over 2,500 special districts now operate in Texas. As of late 2025, Texas local governments carried $368.3 billion in outstanding debt. In the May 2025 election cycle alone, voters were asked to approve $41.3 billion in new principal across 313 propositions.

This is infrastructure at industrial scale. Most of the country does not understand how it works.

The Core Tools

Municipal Utility Districts (MUDs) are the foundational mechanism — independent political subdivisions with the power to issue bonds, levy taxes, and charge user fees. A developer builds infrastructure on raw land, then the MUD issues bonds to reimburse the developer once the TCEQ confirms the infrastructure is complete and a sufficient tax base exists. Bond proceeds repay the developer. Property taxes on homes and land within the district service the debt. Over 950 MUDs now operate in Texas, and every major master-planned community in the Houston metro area — The Woodlands, Clear Lake City, First Colony — was built using them.

The critical structural feature: developers do not get reimbursed until real value exists on the ground. This is reimbursement finance, not speculative finance. The developer takes construction risk. The bond market takes credit risk only after delivery.

Public Improvement Districts (PIDs) are created by cities under Chapter 372 of the Local Government Code. Unlike MUDs, PIDs are not independent governments — the municipality retains control, issues the debt, and administers the district. PID assessments are fixed, several (one owner’s default cannot increase another’s obligation), and can be prepaid. PIDs finance a broader range of improvements than MUDs: parks, trails, lighting, pedestrian infrastructure, affordable housing, and mass transit facilities. The tradeoff is administrative complexity and ongoing municipal oversight.

Tax Increment Reinvestment Zones (TIRZs) are the Texas version of tax increment financing. A municipality freezes the property tax base within a zone, then captures the incremental revenue generated by rising property values and reinvests it within the zone. TIRZs fund streetscaping, utilities, transit stations, and public facilities. They do not create new taxes — they redirect existing revenue. The risk is that the city foregoes increment revenue for decades, betting that the improvements will eventually produce a stronger tax base.

Water Control and Improvement Districts (WCIDs) are older cousins of MUDs with broader authority, including the power to purchase existing water facilities. Management Districts fund supplemental services — security, landscaping, marketing — for commercial areas.

What Makes It Work

Three structural features distinguish the Texas model.

First, infrastructure cost is allocated to the people who create the demand. MUD taxes and PID assessments are paid by property owners within the district, not the broader tax base. This eliminates the most powerful political barrier to growth: existing residents objecting that their taxes will subsidize newcomers.

Second, the system allows infrastructure to be built in advance of municipal capacity. Texas cities grew faster than their governments could accommodate. Special districts bypassed constitutional debt limits and political resistance by creating new, purpose-limited governments with their own tax base.

Third, phasing discipline is enforced. The TCEQ will not approve MUD bonds until infrastructure is complete and a tax base is established. The Attorney General must approve all MUD bonds. PID assessments require city council authorization. These sequential gates prevent bond issuance from running ahead of physical reality.

What the Rest of the Country Should Learn

Any state serious about housing production should study these tools and adopt the structural discipline while fixing the governance weaknesses.

Separate infrastructure finance from general municipal politics. Let beneficiaries fund their own infrastructure. This removes the political barrier to growth.

Use reimbursement structures, not speculative ones. Require infrastructure delivery before bond issuance. The discipline of building first is what makes the credit work.

Allow multiple tools for multiple problems. MUDs for utilities, PIDs for amenities under municipal control, TIRZs for redevelopment. One mechanism cannot serve every purpose.

Set hard timelines for governance transition. Developer control during formation is pragmatically necessary but must be time-limited.

Build transparency infrastructure before physical infrastructure. Texas passed HB 103 in 2025 creating a centralized bond transparency database. This reform should have come decades earlier — and should be a precondition, not an afterthought, in any state adopting these tools.

Pair finance tools with planning standards. The financing mechanism is agnostic about urban form. It will fund good places and bad ones with equal efficiency. Without design codes, connectivity requirements, and public space obligations, the system builds sprawl with better plumbing.

Texas solved the capital problem. But capital without governance is extraction. Infrastructure without planning is sprawl. The question for every other state is not whether to build the machine — it is whether to build the constraints before the machine starts running.

In April 2026, Toll Brothers signed a deal to acquire substantially all the assets of Buffington Homes of Arkansas, the largest luxury homebuilder in Northwest Arkansas. Buffington controls more than 1,500 lots across nine communities in the Fayetteville-Bentonville corridor, building first-time and move-up homes priced from $400,000 to over $1 million, with average selling prices around $650,000. Founding partners Clay Carlton and Mike Lamberth will stay on to lead land acquisition and development. This is Toll Brothers’ sixteenth homebuilder acquisition since 1995. Financial terms were not disclosed.

The deal is not surprising. What is surprising is that it took this long.

Northwest Arkansas has been quietly assembling the conditions that attract national capital for more than a decade. The metro has crossed 600,000 residents and adds roughly 250 new people per week. Three Fortune 500 companies — Walmart, Tyson Foods, and J.B. Hunt — anchor the economy, drawing nearly 1,300 suppliers and vendors to the region. Walmart’s new 350-acre mass-timber headquarters campus opened in Bentonville in 2025, housing 15,000 employees in twelve buildings designed into the city’s street grid, with the 40-mile Razorback Greenway running through the center. Home prices in Benton County averaged $471,427 as of late 2025, a 60.6% increase over five years. Residential building permits hit 3,000 in 2024, the highest since 2006.

Toll Brothers’ business model is built to exploit exactly this profile. The company serves move-up and empty-nester buyers — over 70% of its business — with a national average sale price near $976,000. Its buyers are less sensitive to mortgage rates than entry-level purchasers. They spend roughly $206,000 per home on design studio customizations, structural options, and lot premiums. CEO Karl Mistry has described the company’s strategy as entering markets where affluence is established or rising, partnering with strong local operators, and scaling from a position of alignment. Buffington’s $650,000 average sale price, strong margins, and local land position fit that playbook precisely. So does a region where area median income for a family of four recently crossed $101,800.

The question now is where the growth goes. The four core cities — Bentonville, Fayetteville, Rogers, and Springdale — are increasingly built out. Sixty-one percent of regional population growth is already occurring outside those cities. The next wave of residential development is pushing into western Benton County — Centerton, Cave Springs, Elm Springs, Tontitown, Highfill — where land is available but infrastructure is strained. Sewer capacity is a binding constraint; municipalities have spent years working to connect to the Northwest Arkansas Conservation Authority’s regional wastewater treatment plant. Highway 112 corridor improvements will link Fayetteville through Tontitown to Bentonville, opening new residential and commercial frontage. A proposed north-south connector between the Bella Vista Bypass and U.S. 412 could unlock large parcels near the regional airport in Highfill, one of the few communities with significant undeveloped land remaining in the county.

Toll Brothers’ entry accelerates a pattern already underway. With Buffington’s 1,500 controlled lots, an in-house architectural, engineering, mortgage, and land development operation, and the capital to acquire additional positions, Toll Brothers can move faster along these growth corridors than a privately held regional builder could alone. The retention of Carlton and Lamberth — the local operators who know which land to buy, how to phase, and which product fits which submarket — is the piece that makes the strategy executable rather than theoretical.

Whether Toll Brothers reshapes Northwest Arkansas or simply scales what Buffington already built will depend on what happens at the edges: whether the outlying towns can deliver infrastructure fast enough to support the lots, whether the region develops coordinated land-use policy across its patchwork of jurisdictions, and whether the design standards that have distinguished this metro — Crystal Bridges, the Greenway, Walmart’s campus — hold as production volume increases. The capital has arrived. The land position is secured. What remains to be built is the governance to match the growth.

The last couple of weeks have been a whirlwind. An important shift in monetary policy, namely Basel III, could be happening soon. To provide some context, let’s go to 2008.The Global Financial Crisis (GFC) was not merely a crisis of distressed assets; it represented a systemic collapse of confidence across the entire capital market. The credibility of the securitizations, particularly for mortgage-backed securities (MBS), evaporated, leading to an immediate liquidity freeze. Overleveraged and undercapitalized banks entered a protracted survival mode, swiftly curtailing construction and development (C&D) lending. This contraction was a function of immediate capital preservation and risk reduction requirements, not a granular assessment of project viability. The credit system broadly ceased function, failing to differentiate solvent from insolvent undertakings.1

The ensuing funding vacuum necessitated the emergence of private credit, which crystallized not as a strategic preference but as a market essential. Debt funds, family offices, and opportunistic institutional capital mobilized to supply financing where traditional banks could not. These new entrants fundamentally restructured the lending paradigm: financing became shorter-duration, higher-cost, and buttressed by more stringent structural protections. Whereas conventional banks traditionally prioritized relationship underwriting and supporting long-term development pipelines, private lenders adopted a yield-driven approach, focusing intently on individual transactions. Consequently, residential development finance transitioned from a system predicated on long-standing relationships to one governed by immediate yield metrics.1

This regulatory retreat was formalized by Basel III. Regulators imposed significantly higher capital requirements on what were deemed riskier assets, specifically encompassing construction, land, and development exposures. The resultant impact was purely quantitative, not philosophical. The elevated capital burden drastically diminished the return on equity (RoE) for these specific loan categories, rendering them financially unattractive compared to less risky portfolio alternatives. Banks were thus structurally incentivized to withdraw from the segment of the market that demands the greatest patience and tolerance for initial risk.1

The proposed 2026 re-proposal, last month, constitutes a measured regulatory recalibration. Regulators appear to be acknowledging that certain elements of the post-crisis framework became unduly complicated and, in some instances, decoupled from actual underlying risk. The current proposal seeks to simplify the calculation of capital requirements and alleviate several of the binding constraints that had inflated risk-weighted assets. Crucially, the initiative explicitly aims to repatriate key lending activities—notably residential mortgages and middle-market credit—back onto the balance sheets of depository institutions.1

Basel III Re-Proposal – Simplified Tiering Impact (U.S.)

• Global Systemically Important Banks (GSIBs) / Largest Banks ($250B+ assets)
  Capital rules are simplified but still binding; modest relief, primarily improves capital efficiency on low-risk assets (mortgages, securitization exposure), limited re-entry into construction risk.

• Large Regional Banks ($100B–$250B assets)
  Most meaningful change; reduced capital intensity and less punitive risk-weighting restores lending capacity, especially for residential and later-stage construction—primary driver of spread compression.

• Mid-Sized Banks ($50B–$100B assets)
  Incremental relief with simpler frameworks; moderate increase in balance sheet flexibility, selective re-engagement in local development lending.

• Community Banks (<$50B assets)
  Minimal direct impact; largely unchanged capital treatment, continue relationship-based lending but remain constrained by scale rather than regulation.

For banks, this shift signals not an impending aggressive expansion, but rather the restoration of essential lending optionality. A reduction in the required capital per loan directly enhances RoE, making specific credit categories, particularly residential and later-stage construction financing, economically viable once more. Regional banks, having been most significantly constrained by the preceding framework, are poised to be the crucial swing factor. Their previous limitation was one of balance sheet capacity, not underlying market demand.1

The market dynamic over the ensuing 12 to 24 months is expected to manifest first in pricing adjustments before impacting overall volume. Banks will re-enter the market judiciously, directing competition toward robust deals and established sponsors. Private credit will not be supplanted, but its pricing premium will be attenuated. This heightened competition among balance sheets for the same assets will drive a compression of construction and mortgage spreads, independent of any sudden capital abundance.

Artificial intelligence and robotics are increasingly framed as the macroeconomic release valve: raise productivity enough and central banks can loosen policy without inflation reappearing. The logic is appealing. Productivity—defined as the amount of output produced per unit of input—sets the economy’s effective speed limit. When productivity rises, the economy can grow faster without pushing prices higher, giving monetary policy room to breathe.

History suggests that this relationship is real—but conditional. Productivity only buys policy flexibility when it improves how the system allocates capital, labor, and time. When it does not, easy money merely stretches the clock.

Over the past 50 years, there are three periods where productivity meaningfully altered the constraints facing U.S. monetary policy. Each one matters for understanding how AI might reshape the economy today.

Lesson One: The 1970s — When Productivity Per Worker Slowed and Policy Lost Control

The 1970s are often explained as a story of oil shocks and inflation. That explanation is incomplete. The deeper issue was a sustained slowdown in productivity per worker—output per hour worked.

Several forces converged. Energy prices spiked, but the capital stock was built for an era of cheap fuel. Heavy industry relied on large, centralized plants that could not be easily retooled. Manufacturing processes were rigid and slow to adapt. At the same time, labor force participation rose sharply as baby boomers and women entered the workforce. While positive in the long run, this influx reduced average output per worker in the short run as inexperienced labor diluted measured productivity. Regulation expanded in transportation, energy, and manufacturing, adding compliance costs without raising output.

Consider a steel mill operating with outdated blast furnaces designed for cheap energy. When energy prices rose, the mill’s cost per ton increased sharply. If demand improved and the Federal Reserve attempted to stimulate growth by easing policy, the mill could not efficiently increase output. Each additional unit of production required disproportionately higher input costs. The result was higher prices, not higher volumes.

This is why the Fed ultimately raised rates aggressively. With productivity per worker weak, easing policy translated directly into inflation. Rate hikes were not a choice; they were a constraint. Monetary policy had lost flexibility because the supply side of the economy could not respond.

Lesson: When worker-level productivity is weak, economic growth translates into higher prices rather than higher output, leaving monetary easing to amplify inflation instead of expanding production.

As productivity per worker stalled and inflation accelerated, the Federal Reserve raised the Federal Funds rate from roughly 5–7% in the mid-1970s to a peak near 20% in 1981, using restrictive policy to suppress inflation when supply could not expand.

Lesson Two: The 1990s — Productivity Forces Capital to Move

The late 1990s demonstrated the opposite dynamic. Information technology raised productivity not by squeezing labor harder, but by reorganizing how firms operated. Real-time inventory systems, enterprise software, and improved logistics reduced waste, shortened production cycles, and lowered coordination costs. Output per worker rose meaningfully.

Crucially, productivity gains forced reallocation—the movement of capital away from lower-productivity uses toward higher-productivity ones.

To understand this, it helps to define marginal capital. Marginal capital refers to assets or investments that generate returns just sufficient to cover their cost of capital. In healthy systems, marginal capital is fragile. When conditions change, it must either improve, restructure, or exit.

In the 1990s, retailers and manufacturers that relied on bloated inventories, inefficient warehousing, and slow distribution became uncompetitive. Capital tied up in excess storage and redundant labor could no longer justify itself. Firms that failed to modernize exited or restructured. That capital was redeployed into automated distribution centers, scalable software systems, and firms capable of operating with far higher efficiency.

This mattered for monetary policy. Productivity gains lowered unit costs across the economy. Wages could rise and unemployment could fall without igniting inflation because inefficient uses of capital were being cleared out. Price signals remained sharp. The economy’s supply capacity expanded.

Lesson: When productivity gains are broad and force reallocation, monetary policy gains flexibility. Growth accelerates, inflation stays contained, and capital moves toward higher-value uses.

During the technology-driven productivity boom, the Federal Reserve maintained a relatively moderate rate range—generally 3% to 6.5% between 1992 and 2000—allowing strong growth and falling unemployment without triggering inflation.

Lesson Three: The 2010s — When Capital Survived Instead of Moving

After the Global Financial Crisis, monetary policy remained extraordinarily loose for more than a decade. Interest rates approached zero and central bank balance sheets expanded dramatically. Inflation stayed low, leading many to assume productivity had quietly improved.

It had not.

Productivity growth during the 2010s was weak. The defining feature of the period was not efficiency, but survival.

Cheap financing allowed marginal capital to persist indefinitely. Instead of exiting, weak businesses refinanced. Instead of being written down, underperforming assets were extended. This suppressed price discovery—the process by which markets determine value through transactions that force repricing, losses, and exits.

Price discovery requires clearing. Assets must trade. Prices must fall when demand weakens. Capital must be written down when it no longer earns its keep. Post-2008 policy softened this mechanism. Sellers waited. Buyers delayed. Transactions slowed. Capital stayed in place.

A commercial building with declining tenant demand, for example, did not need to clear at a lower price if financing costs were negligible. It survived—not because it was productive, but because it could endure.

The result was stability without throughput. Capital survived, but it did not redeploy. Reallocation slowed. Fewer decisions were forced. Inflation remained low not because productivity surged, but because demand was constrained and gains were unevenly distributed.

Lesson: Loose monetary policy without productivity preserves marginal capital, suppresses price discovery, and slows the economy’s internal circulation.

Following the financial crisis, the Federal Reserve cut the Federal Funds rate from 5.25% in 2007 to 0–0.25% by 2008, holding it near zero for most of the next decade despite weak productivity growth to preserve financial stability and asset markets.

AI and the Productivity Question Today

AI complicates this history because it can raise labor productivity without improving total factor productivity (TFP).

• Labor productivity measures output per hour worked. A firm becomes more labor-productive if the same output is produced with fewer workers or fewer hours.

• Total factor productivity measures how efficiently the entire system converts labor, capital, and technology into output. It improves only when coordination, allocation, and processes improve across the economy.

AI is extremely effective at raising labor productivity. A customer service team using AI can handle the same call volume with half the staff. A legal department can process contracts faster with fewer associates. Output per worker rises.

But total factor productivity may not.

Consider a logistics firm that uses AI to optimize routing and reduce labor costs, but operates in a market with limited competition and protected pricing. Costs fall internally, but prices do not. Volumes do not rise. Capital does not redeploy elsewhere. The firm’s margins improve, but the system does not become more efficient. Productivity shows up inside the firm, not in the broader economy.

If markets remain distorted—through market power, policy protection, or delayed clearing—AI compresses costs without improving allocation. Inflation may stay low, but throughput does not increase.

This is the central risk: repeating the 2010s dynamic with better tools.

Jobs, Housing, and the Risk Ahead

The labor market implications hinge on the difference between mass unemployment and underutilization.

Mass unemployment is visible and abrupt. Large numbers of workers lose jobs at once, demand collapses quickly, and policy responds aggressively.

Underutilization is quieter. Workers remain employed, but:

• work fewer hours,

• accept lower-quality jobs,

• cycle between roles,

• experience income volatility,

• operate below their skill level.

AI is more likely to produce underutilization than mass unemployment. Tasks disappear, roles thin out, and middle layers compress. Employment persists, but income growth weakens. Demand softens gradually rather than collapsing. Monetary policy can mask this with easier conditions, but it cannot fully replace lost labor income.

Housing is where these dynamics converge. AI can lower construction and design costs, but if land constraints, zoning delays, and financing structures prevent clearing, productivity gains will be capitalized into asset values rather than affordability. At the same time, a labor market characterized by underutilization produces fewer households with stable, mortgage-ready income. Demand shifts toward rentals, while ownership concentrates among those with durable balance sheets.

The central risk is not collapse. It is a low-speed equilibrium: higher task efficiency, slower capital movement, and stagnant system-level productivity. Whether AI restores throughput or simply perfects delay will depend less on the technology itself than on whether markets are allowed to clear and capital is forced to move.

In 2025, the housing market resembled a diner kitchen at 2 a.m.—undersupplied, overheated, and full of improvisation. Builders worked the griddle with one eye on rate sheets and the other on a thinning margin. The Fed finally turned down the heat in Q4, and demand, like a late-night crowd, stumbled back in—still groggy, still price-sensitive, but hungry enough to bite. Resale inventory stayed locked in the walk-in freezer. New homes came out smaller, cheaper, and with free financing on the side. It wasn’t pretty, but somehow, the orders kept going out.

Revisiting 2025 Themes

Q1: Margin Resilience and Fed Head-Fakes Homebuilders began the year defying gravity. Margins stayed strong despite rate volatility. By Q4, gravity won. Margins compressed. NVR’s dropped from 23.4% to 21.0% YoY. Incentives soared. The Fed finally cut rates in December (0.25%), confirming the pivot we anticipated. The year opened with ambiguity and closed with policy clarity.

Q2: Frozen Supply and Bifurcated Demand Listings stayed scarce. Existing inventory ended the year ~12% below pre-pandemic levels. Meanwhile, builders undercut the resale market: new homes sold for ~$20,000 less than existing ones. The demand split widened. Builders scaled prices down and leaned into incentives. The resale market stayed stubborn.

Q3: Hints of Loosening The Fed’s shift materialized. The December rate cut followed two earlier ones, easing the 30-year mortgage rate from above 7.5% down to ~6%. Builder sentiment (NAHB Index) rose to 37 in October from 32 in September. Still negative, but trending positive. We forecasted a pivot. It landed.

Homebuilder Scorecard

Industry Trends Volumes held. Margins shrank. Incentives were the lever. Builders continued land conservatism. Speculative building was minimal. Starts aligned with demand.

D.R. Horton (DHI) Closed 23,368 homes in Q4. Revenues: $8.5B. Orders lagged at 20,078 units, signaling backlog reliance. Affordable segments drove volume. Incentives were standard; rate buydowns commonplace. Gross margin dipped to ~20.5%. Lot control remained disciplined (25% owned, rest optioned).

Lennar (LEN) Pulled back slightly in Q4 after Q3’s volume push. Targeted ~22,000 closings. Operating margin fell to 17.5% from 22.1% YoY. Incentives expanded; 38% of builders nationally cut prices (~6% average cut). Lennar emphasized land-light operations, preferring flexibility over inventory risk.

NVR Orders down 16% YoY in its reported Q3. Cancellations rose to 19%. Gross margin fell to ~21%. Land deposit impairments surfaced. Still profitable, still cautious. Q4 deliveries estimated in low 5,000s. Earnings met expectations despite cooling demand.

The Fed’s Pivot

Three rate cuts in 2025, finalizing at 3.5–3.75%. Mortgage rates fell from mid-7% to low 6%. A $500K mortgage at 6% saved ~$600/month versus 7.5%. Demand picked up. Model home traffic improved. Mortgage apps climbed in December. Builder confidence responded.

The Fed shifted tone: inflation easing, unemployment rising modestly (4.4% by Q4), rate hikes paused. Market response was swift. 10-year Treasury yield fell to ~4.1%. Capital opened back up. MBS spreads narrowed. Homebuilder stocks rallied ~15% off midyear lows.

Capital Markets

Affordability and Credit Affordability improved marginally. Income gains and rate drops helped. FHA delinquencies crept up. Lending standards stayed tight. New programs (e.g., 2-1 buydowns) gained traction. Some assumable loan strategies resurfaced.

Investor Activity Bulk home sales to SFR investors continued. Builders leaned on institutions to clear spec. iBuyers stayed muted. Selective investor re-entry seen in discounted markets.

Builder Stocks S&P Homebuilders Index closed up ~5% for the year. Valuations remained below market average. Capital confidence returned ahead of fundamentals.

Regional Snapshots

Inventory Divergence

• South/West: inventory above pre-COVID levels (West +3%, South +5.7%). Builders overdelivered.

• Northeast/Midwest: inventory well below norms (Midwest -33%, Northeast -48%). Scarcity preserved price.

Refuge Markets Cleveland, Milwaukee, Pittsburgh, Providence led price appreciation. Prices rose >20% since 2022 in some. Affordable metros drew price-sensitive buyers priced out of coastal or Sunbelt markets.

Sunbelt Cooldown Austin: median price -4% YoY. Las Vegas: inventory +33% YoY. Atlanta: inventory +13%. Builders heavily incentivized. Correction was mild, not distressed. Owners held, not dumped.

Coastal Metros NYC, LA, Boston saw stable demand and low resale listings. San Francisco lagged due to soft tech employment. Miami stayed strong, buoyed by cash buyers and in-migration.

Conclusion

2025 began with ambiguity and margin strength, passed through limbo and demand bifurcation, and ended with a monetary pivot and early signs of re-engagement. Builders gave up margin but kept pace. The Fed shifted posture. Inventory remained fractured regionally.

The market didn’t crash. It adjusted. Supply was frozen, but builders improvised. Rates peaked, then fell. Confidence rose, tentatively. The resilience of demand persisted in new forms: smaller homes, cheaper markets, creative financing.

If 2023 was shock, and 2024 was stall, then 2025 was recalibration. Not yet recovery, but something sturdier than survival. The coming year has fewer headwinds. That alone is a change.

Every developer eventually meets the same problem: the deal looks great on paper until the city gets involved. A pair of lots, an ideal zoning tag, a plan for six units that fits the code—until someone in Planning discovers a gap in the chain of title or an old subdivision that was never properly recorded.

That was the situation on a recent Glendale infill deal: two R3-zoned parcels totaling roughly 7,500 square feet. Flat, centrally located, and perfectly sized for a small six-unit multifamily. The only problem—City staff couldn’t confirm the lots were legally recognized as separate or buildable together.

The seller wanted to close. The buyer wanted certainty. The City wanted 90 days for a Certificate of Compliance (CoC) review. Everyone’s clock was ticking.

1 · Start With Title Reality

Before debating design or financing, clarify ownership structure. The fastest way to lose six months is to skip a title deep dive. Many LA-area parcels were split and resold in the 1980s and 1990s without formal lot ties or certificates. That’s the source of most “ghost parcels” that appear buildable but aren’t legally recognized.

If you find irregularities, one fix is a lot tie and resubdivision: joining the parcels under a new map and then splitting them again cleanly. It’s tedious but decisive, and it protects future entitlements and financing from legal challenges. Most lenders will insist on this step before closing construction debt.

2 · Know If You’re Ministerial or Discretionary

Under LA’s updated housing laws, a six-unit R3 project that meets objective standards should qualify as ministerial—no hearings, no discretionary review, just plan check. But “should” is the key word. Cities often apply internal guidelines inconsistently when older parcels are involved.

The safest path is to build your submittal package around objective metrics—height, FAR, open space, and parking—and document compliance meticulously. The more you can prove your project fits within ministerial boundaries, the harder it becomes for staff to shift it into discretionary review later.

3 · Bridge the Uncertainty With Seller Paper

When the city’s review horizon doesn’t match your closing schedule, the cheapest source of capital is the person across the table: the seller.
A seller carryback lets both parties move forward while sharing the risk of entitlement delay. The structure is straightforward:

• Seller holds a short-term note (12–24 months) at a negotiated rate.

• Buyer releases deposits based on city milestones (intake accepted, corrections received, CoC cleared).

• The final payment triggers upon approval or recorded map.

This approach keeps the property out of the hard-money market, which often charges 12–14% and limits leverage to 50–60% of land value. For early-stage entitlement risk, seller paper is almost always cheaper and better aligned.

4 · Package for Permits, Not Entitlements

If the site meets objective standards, focus on documentation that satisfies Building & Safety, not on discretionary-level renderings. Keep consultants narrow: architect, civil engineer, and legal. Skip the community outreach consultants and public affairs spend until you confirm you actually need it. Time and cost discipline during uncertainty are survival tactics.

5 · Manage Expectations on Value and Carry

Even if the project is technically “by-right,” the gray area around lot recognition or compliance will lower land leverage. Underwrite 50–60% loan-to-value, 12–14% interest, and a 24-month term as your base case. If you can beat that with seller terms, consider it a win.

Remember: capital prefers certainty, not creativity. When title or zoning questions exist, the developer must own the patience premium.

Conclusion — The Cost of Clarity

Deals like this test a developer’s discipline more than their imagination. The right move isn’t to over-engineer the capital stack—it’s to remove ambiguity at the source. A clean title, a ministerial path, and a seller aligned on timing can turn an unfinanceable property into a quiet, successful build.

Entitlements can be unpredictable, but clarity is a form of collateral. The more certainty you can create on paper, the faster the capital will follow.

The three most common paths for adding units in Los Angeles—lot splits, condo maps, and ADUs—each promise density but deliver it in very different ways. On paper, they all increase housing. In reality, they change everything about how a project is financed, sold, and valued.

Choosing the wrong path early can cost years of time and millions of dollars in capital efficiency. The right one depends on who your buyers are, how much time you can afford, and what the site can actually support under its overlapping rules.

1 · The Market Logic Behind the Decision

In residential development, scale defines liquidity.

• One to four units falls into the consumer mortgage world. Buyers can use FHA, Fannie Mae, or Freddie Mac financing, and banks understand the product. It’s the broadest buyer pool in the housing market.

• Five units or more pushes the project into commercial lending territory. That limits who can buy, often requiring investors who qualify on property income instead of personal credit.

The takeaway is simple: a small building is a consumer product, but once you cross four units, it becomes an investment product. The financing markets—and therefore exit values—follow that line.

Condo mapping, therefore, is often the bridge between the two. It allows a larger building to be split into individually financeable pieces again, each eligible for residential buyers. Done right, it restores liquidity to an otherwise illiquid asset.

2 · When Condo Mapping Works

Mapping shines in high-value submarkets where per-unit prices justify the cost. Five-plus-unit buildings in places like Venice, Silver Lake, or West Hollywood often gain 10–20% in total value once individually mapped. Each unit can sell to a different end buyer, and lenders can size construction loans with multiple take-out options.

However, the benefits taper off on smaller projects. For a two-unit property, the condo-mapping premium may not cover legal, survey, and HOA formation costs. Developers in these cases often achieve higher returns by selling the whole property to one buyer rather than splitting it.

3 · The Lot Split Alternative

SB 9 created a new tool for subdividing single-family lots into two fee-simple parcels. This path offers long-term flexibility—each parcel can be financed, built, or sold separately—but it carries heavier process and time risk. Lot splits typically require two to three years from start to permit. Driveway easements, utility separations, and access issues can eat months and invite neighborhood scrutiny.

For sponsors with patient capital or long-term hold strategies, it’s a powerful wealth-building tool. For merchant builders relying on construction loans, the clock often kills the margin.

4 · ADUs and JADUs — The Quickest Path to Incremental Density

Accessory Dwelling Units (ADUs) and Junior ADUs (JADUs) are the simplest legal way to add income or resale value. They rarely trigger public hearings, avoid subdivision requirements, and benefit from recent state laws that restrict local interference.

But while ADUs are fast, they have appraisal challenges. Most appraisers still value them as accessory space rather than stand-alone homes, which means their cost may exceed their immediate resale contribution. New legislation allowing condo-ization of ADUs could change that, but it’s too early to know how lenders and buyers will treat these sales.

5 · Parking and Appraisal Realities

Each path interacts differently with parking policy. Under AB 2097, projects near major transit stops can reduce or eliminate parking requirements. That relief can make or break feasibility on small infill lots. Yet even when legally exempt, missing parking can reduce sale prices. Developers should underwrite that trade-off carefully.

Conclusion — Matching the Path to the Capital

Each strategy serves a different kind of developer.

• ADUs reward small, nimble owners focused on yield.

• Lot splits favor long-term builders who can absorb time risk.

• Condo maps fit merchant developers chasing liquidity in high-value markets.

The smartest teams begin with the exit market in mind. What kind of buyer will ultimately write the check? Work backward from that answer and you’ll know which path belongs on your site plan.

Private credit has quietly replaced traditional banks as the backbone of commercial real-estate finance. What began as a stopgap during the pandemic has become a structural shift in how projects get capitalized. The change didn’t happen overnight; it grew out of a decade of policy decisions and a new kind of lender appetite that prizes yield and speed over regulation and relationship banking.

Understanding this shift is essential for any developer or sponsor trying to size loans in today’s market. The rules have changed, but the fundamentals—trust, clarity, and collateral—still determine who gets funded.

1 · The End of Cheap Money

From 2008 through 2021, the United States effectively ran on quantitative easing. The Federal Reserve bought trillions in mortgage-backed securities and Treasuries, holding rates artificially low and creating an ocean of cheap liquidity. Underwriting softened because lenders knew they could sell their paper into a hungry secondary market.

That era ended in 2022. The Fed moved to quantitative tightening, allowing roughly $30 billion of assets per month to mature off its balance sheet without replacement. Liquidity began to drain, spreads widened, and the same lenders who once courted every borrower started scrutinizing every deal. Projects that penciled at 3.5% debt suddenly faced double the interest cost.

2 · The Rise of Private Credit

As banks pulled back, private credit filled the gap. Pension funds, asset managers, family offices, and high-net-worth investors discovered they could originate debt directly and earn double-digit yields with asset-backed security. By 2023, non-bank lenders had surpassed traditional banks in total originations across many segments of commercial real estate.

Private credit behaves differently. It doesn’t rely on deposits, doesn’t sell paper to the Fed, and often has more flexible mandates. For developers, that means more creativity—non-recourse structures, interest-only terms, higher leverage—but also higher pricing and tighter covenants.

3 · Underwriting by Economics, Not Personality

Bank lending traditionally relied on borrower strength: liquidity, net worth, and a long relationship history. Private credit looks at something else—the project’s standalone economics.

Two metrics define this new world: Loan-to-Cost (LTC) and Debt Yield. Debt Yield is simply net operating income divided by loan amount, and it has largely replaced DSCR as the sizing constraint. Most lenders now target 6–8% at stabilization. If the yield clears, the deal can work, even without heavy recourse. If it doesn’t, no balance-sheet strength will save it.

That’s why design and entitlement precision now matter as much as financial reputation. A plan that clearly produces collateral and cash flow is easier to finance than one with a vague timeline or unclear exit.

4 · Collateral Is the New Covenant

When debt is expensive, leverage comes from structure, not rate. Developers are increasingly creating additional collateral through lot splits, condo maps, and phased takedowns. These maneuvers give lenders multiple exit paths and, therefore, higher comfort with leverage. A single 10-unit building on one APN may cap at 60% loan-to-value, but ten mapped units can reach 70–75%.

The more distinct assets a lender can sell or refinance separately, the stronger your negotiating position.

5 · The 2026 Outlook

Interest rates may drift lower through 2026, but the shift to private credit is the new normal. Banks will remain cautious; regulators prefer it that way. Sponsors who learn to package projects for private lenders—clear collateral, solid cost control, fast exit visibility—will raise faster and on better terms than those still waiting for bank appetites to return.

Conclusion — The Professional Borrower

The new borrower profile is not just creditworthy but investor-ready. Private lenders want transparency, updated budgets, and an understanding of where value can be unlocked through design and phasing.

For developers, this is liberating. You’re no longer selling your personal balance sheet—you’re selling the clarity of your deal. In today’s market, the spreadsheet is the real guarantor. Master that, and capital will find you.

Los Angeles doesn’t really have zoning in the simple, one-page sense developers expect. It has zoning layered under a tangle of specific plans, community plan implementation overlays (CPIOs), and coastal regulations that often overrule base entitlements. Think of it as a layer cake: the general plan sets the vision, zoning gives a nominal density, but the overlay on top quietly changes the recipe.

If you’ve ever opened the zoning map, seen “R3,” and imagined six or eight units—only to learn you’re limited to two—you’ve experienced this hierarchy firsthand.

1 · Specific Plans & CPIOs — the Neighborhood Constitution

Specific Plans act like miniature constitutions for individual neighborhoods. In theory, they protect character; in practice, they can lock density and height far below what the zoning promises. The Venice Coastal Specific Plan, for example, limits unit counts, height envelopes, and even roof shapes. Developers unfamiliar with those details often spend tens of thousands on design work that cannot be approved.

A CPIO, such as those applied in South LA or Hollywood, adds an extra layer—design standards, setbacks, or step-backs that supersede base zoning. Unless you study them early, they can quietly erase half your yield before you’ve even opened Excel.

2 · The Coastal Zone — Slow-Motion Entitlement

If your parcel touches the California Coastal Zone, you answer not only to the City but to the Coastal Commission. Their hearing calendar moves slowly, environmental reviews multiply, and public-access findings can drag out projects for years. Even ministerial projects can take five to seven years once Coastal review begins. With current capital costs, few deals can carry that long without eroding returns.

3 · CHIP — The Citywide Housing Incentive Program

CHIP is the City’s newest attempt to bring logic to the chaos. It offers density bonuses, parking reductions, and height increases citywide without relying on older TOD overlays. The intent is to help smaller infill developers compete. But the rules are still being interpreted, and few teams have modeled CHIP at full strength. Many are waiting to see how it interacts with existing Specific Plans before committing to those assumptions in a pro forma.

4 · SB 79 and State Pre-emption

Sacramento is rewriting the script at the state level. SB 79 and related laws require cities to approve certain housing types ministerially—especially near transit. In several corridors, the state’s “shall approve” mandate overrides local discretion. The lag between state law and city guidance, however, creates uncertainty for anyone trying to close financing while the rules catch up.

5 · What This Means for Developers and Lenders

Design to the strictest layer first. Build three scenarios—Base Zoning, Overlay, and State Override—and stress-test each for time, soft cost, and carry. Lenders respect a borrower who underwrites real constraints. A zoning memo that explains the overlay stack signals competence and protects credibility when committees or investors review your file.

Conclusion — The Future of the Stack

Los Angeles isn’t one set of rules; it’s a living hierarchy. Specific Plans tighten, state laws loosen, and CHIP tries to stitch the difference. The developers who thrive will be translators—those who can read across city, state, and capital markets to turn friction into feasibility.

Zoning in LA isn’t just about what’s allowed on paper. It’s about understanding how each layer of regulation reshapes value, time, and risk. Master the layers, and you master the city.

Every major infill project eventually collides with the same question: How much density is too much? It’s not just a planning debate—it’s an economic one. The tension between what’s allowed, what’s financed, and what’s politically survivable defines the rhythm of urban development in Los Angeles.

Over the last few years, community resistance has grown more vocal, but paradoxically, so have the returns in those same contested areas. The pattern is almost mathematical: where the neighborhood fights hardest, land values are highest. For developers who can navigate that friction without alienating the community, those are often the projects that outperform.

1 · Resistance as a Market Signal

Neighborhood opposition isn’t random—it’s a proxy for desirability. Areas that attract pushback tend to be walkable, supply-constrained, and near strong amenities. People fight density when they already love the place.

That correlation is critical for developers and lenders alike. A site that provokes organized resistance is almost always a site where finished product values will hold up, even in soft markets. The trick is to distinguish between temporary political turbulence and structural risk. Organized neighborhood groups can delay a project, but they rarely erase demand.

The best sponsors treat community tension as due diligence data: a leading indicator that a market still has heat.

2 · The True Cost of Underground Parking

Most public opposition frames around parking—how much is lost, where it will go, and who will get to use it. But the real issue isn’t public convenience; it’s project economics. Underground parking can destroy a pro forma faster than any policy change.

Each stall can cost $80,000–$120,000 to build in Los Angeles, and that number climbs in coastal submarkets with high water tables. Pumping, shoring, and dewatering add risk that’s almost impossible to quantify early. Many well-intentioned projects die not because of design or demand, but because underground parking doubled in cost before permits cleared.

That’s why new state rules like AB 2097, which remove parking minimums near transit, are so important. They don’t just encourage density—they prevent projects from collapsing under their own excavation.

3 · Designing for Cities, Not Lots

Density becomes sustainable only when design responds to context. Developers who plan like urbanists rather than lot maximizers tend to win both financially and politically. That means aligning building massing with mobility patterns, creating shared spaces that double as visual buffers, and anticipating how tenants will actually move through the neighborhood.

A well-designed mixed-mobility site—bikeable, walkable, near transit—can command premium rents even with fewer parking stalls. In practice, the urban format is becoming the financial format. Lenders have started to see design-conscious planning as a form of risk reduction, not a luxury.

4 · Using Data to Defuse Community Fear

When neighbors show up angry about parking or traffic, numbers often calm the conversation. Show realistic trip generation, modal splits, and peak-hour volumes rather than abstract counts. Back your case with verified ITE data and real traffic studies. Most residents don’t oppose development itself—they oppose uncertainty.

Transparency builds credibility, and credibility shortens entitlement timelines.

5 · The Long Game of Urban Value

Los Angeles is slowly relearning what older cities already know: density, done well, is durable value. Compact, connected neighborhoods age better, appreciate faster, and support stronger local economies. Projects that push that transition forward may be controversial in the moment, but they often define the next decade of pricing.

The developers who consistently win in LA aren’t necessarily the loudest—they’re the ones who can translate friction into design, turn opposition into context, and keep moving when everyone else waits for permission.

Conclusion — Building Where the Heat Is

The most profitable infill isn’t always the easiest; it’s the one that sits on the fault line between aspiration and anxiety. High-friction neighborhoods are where zoning, market, and politics collide—and where lasting value is created.

The goal isn’t to avoid conflict but to understand what it’s telling you. When people fight for their neighborhood, they’re also affirming its worth. A developer who listens to that signal, refines the design, and still moves forward doesn’t just build housing. They build the next chapter of the city.