Glossary · Data provenance
Plain-English definitions, source provenance, and honest limitations for the Holland Market Research desk. This page is meant to keep the product source-first: every metric, score, forecast, and caveat should be understandable before it is used in a market discussion.
One-line definitions, computation notes, and reading rules for the core measures.
Computed: Annualized apartment inventory / delivery growth from the source series for each metro and scenario where available.
Read it as: Positive supply growth means new competitive units are arriving; high supply without matching demand raises lease-up risk.
Computed: Apartment demand growth from Green Street history/forecast, with public employment and migration features reserved for the long-history challenger panel.
Read it as: Demand growth is the occupancy/rent support side of the equation, not a recommendation by itself.
Computed: Demand growth minus supply growth over the same period, with frequency and source vintage retained in lineage.
Read it as: Positive balance indicates absorption is outrunning new supply; negative balance flags overbuild pressure.
Computed: Running sum of quarterly demand-minus-supply gaps from the available history window.
Read it as: Shows whether a market has accumulated undersupply or oversupply across cycles, not just in the latest quarter.
Computed: Weighted blend of four pillars: undersupply, forward demand, inverse supply threat, and valuation, using the documented Phase 2c weights.
Read it as: A research prioritization tool for comparison, not a public ranking or investment instruction.
Computed: Undersupply captures balance/cumulative gap; forward demand captures scenario demand; supply threat penalizes future deliveries; valuation captures cap-rate / pricing context.
Read it as: Use pillar disagreement to see why a market ranks where it does instead of relying on a single blended number.
Computed: Green Street 2026Q1 scenario paths: Baseline, stronger growth variants, moderate recession, and protracted slump.
Read it as: Scenario lines show sensitivity to macro paths; they are not new vintages or realized outcomes.
Computed: Walk-forward challenger model distribution with p10 / p50 / p90 bands, currently cross-checking Green Street rent growth rather than replacing it.
Read it as: Use it as disagreement evidence and uncertainty framing; challenger = cross-check, not ground truth.
Computed: Market asking rent per unit from the approved private July 2026 CoStar/Linneman workbook, quarterly market/MSA grain.
Read it as: Advertised rent before concessions; shown as governed descriptive history on the Rent History panel only.
Computed: Market effective rent per unit net of concessions, from the same private workbook (available 2000 Q1 onward for most markets).
Read it as: What a lease actually achieves; read beside asking rent to see concession pressure across cycles.
Computed: Nominal values are workbook dollars; real growth is trailing-12-month nominal growth deflated by national all-items CPI (CPIAUCSL).
Read it as: Real growth strips inflation so long histories compare across eras; it is descriptive, not predictive.
Computed: Share of variance in market real asking-rent growth described by the U.S. series over 2001 Q1–2026 Q1, adjusted for the single regressor. The product label is always Adjusted R², never R-Squared.
Read it as: Below 0.50 the breakeven reliability gate fails closed and no precise threshold is shown.
Computed: Intercept of the contemporaneous regression of market real asking-rent growth on U.S. real asking-rent growth, displayed in percentage points.
Read it as: Descriptive same-quarter sensitivity; never evidence of skill or a predictive edge.
Computed: Slope of the same contemporaneous regression, with HAC four-lag uncertainty and rolling/subperiod sign-stability checks.
Read it as: How strongly the market co-moves with the national series in the same quarter; descriptive, not causal.
Computed: Minus alpha over beta: the U.S. real asking-rent growth at which expected same-quarter market real growth is zero. Rendered only when the balanced fail-closed gate passes: Adjusted R² ≥ 0.50, positive HAC beta interval, beta-sign stability, and a bounded Fieller interval no wider than 2.00 pp.
Read it as: When the gate fails the panel names the governed reasons and shows no precise threshold.
Computed: Heteroskedasticity-and-autocorrelation-consistent (Newey-West) covariance with four lags and finite-sample t inference behind the alpha/beta intervals.
Read it as: Quarterly rent series are autocorrelated; HAC keeps the stated uncertainty honest.
Computed: The era-comparison benchmark: the simple average of all 41 analyzed workbook metros, each weighted equally.
Read it as: A market's premium is measured against this benchmark in basis points; it is context, not a target.
Computed: Published windows use each measure's full governed availability; aligned windows start both measures at the same quarter so asking and effective compare over one period.
Read it as: Window choice changes means; the panel derives its narrative from governed era geometry rather than hardcoded dates.
Sources actually used today. License field intentionally distinguishes free-public from licensed/current private inputs.
| Source | What it provides | Geography | Earliest date | Frequency | License | Role | Known limitations |
|---|---|---|---|---|---|---|---|
| Green Street forecast | Apartment supply, demand, rent growth scenarios, fundamentals and valuation inputs | Covered apartment markets | Varies by metric; current site uses one 2026Q1 vintage | Quarterly / scenario | current source delivery | primary | Green Street forecast = single 2026Q1 vintage; accuracy is unscoreable until later vintages/realizations exist. |
| Green Street supply / fundamentals / permits files | Supply growth, permits, cap rates, CPPI, market fundamentals | Metro / Green Street market | Source-dependent | Quarterly / annual source-dependent | current source delivery | primary | Coverage differs by market and metric; missing or sparse coverage is flagged rather than silently filled. |
| CoStar rent history / Linneman Associates workbook | Quarterly asking/effective rent history; TenureOS derives real growth, alpha/beta, Adjusted R², and breakeven from it (recipient-derived analysis, Holland/TenureOS) | Quarterly market/MSA; 41 analyzed metros (35 current studies + 6 rent-history-only) | Source rows 1982 Q1; asking from 1986 Q3 | Quarterly | approved private July 2026 delivery — Holland internal only; no public redistribution; no raw workbook download | evidence sidecar | Underlying data is CoStar; workbook/analysis is Linneman Associates; derivations are TenureOS, distinguished from source facts. Shown separately from Green Street and never blended. |
| BLS CPI — FRED CPIAUCSL (SA) | National all-items CPI used to deflate nominal rent growth into real growth for the Rent History panel | National | 1982 Q1 quarterly averages | Quarterly (from monthly) | free-public, verified against the workbook CPI cache | deflator | 176 shared verified quarters against the workbook cache; the 2025 Q4 value is an exception_missing_official_month treatment and is not publicly reproduced. |
| Linneman page-1 metro CPI-rent analysis | Metro CPI-vs-rent PDF exhibit from the same July 2026 delivery | Metro | n/a | One-time exhibit | approved private July 2026 delivery | deferred evidence | Explicitly deferred: the required structured source layer for the page-1 metro CPI-rent analysis is unavailable, so no product functionality is built on it. |
| CM Alert CMBS + CRE CLO databases | Securitized debt issuance and multifamily share context | National only | Source-dependent | Quarterly aggregate | current source delivery | context | debt data = national-only (never metro); deal files are not joined to individual markets. |
| FRED / Census BPS / BLS / HUD / ACS / IRS migration | Planned free-public long-history permits, employment, wages, FMR, income, population, household formation, migration features | Metro / CBSA / county reconciled through market_crosswalk.csv | Target late-1980s to early-1990s where available | Monthly / quarterly / annual native frequencies | free-public | planned primary/context | Phase 4b only; annual series are never fabricated to monthly, and unmatched metros must be reported. |
| Zillow | ZORI multifamily and ZHVI recent rent/value breadth | Metro / Zillow geography | ZORI ~2015; ZHVI ~2000 | Monthly | free-public CSV | validation | Zillow/Apartment List = shallow (2015/2017+); useful for validation and recent breadth, not deep cycle training. |
| Apartment List | Recent rent and vacancy measures | Market / metro | ~2017 | Monthly | free-public CSV | validation | Shallow history only; not a substitute for licensed apartment-specific absorption history. |
| Challenger model | Rent-growth p10 / p50 / p90 fan and disagreement checks | Covered markets with feature history | Feature-window dependent | Quarterly model panel | internal model output | cross-check | challenger = cross-check, not ground truth; with more true apartment cycles the current AR(4) verdict could flip. |
Not pulled, not scraped, and not assumed. These are future accuracy upgrades only if team licenses are confirmed.
Status: general/live access not-yet-licensed; do not scrape or ingest. The approved private July 2026 Linneman Associates workbook delivery (CoStar-derived) powers the Rent History evidence panel only; that approval does not extend to general or live CoStar access.
Unit-level supply, construction, leasing, absorption, and rent evidence across broader apartment coverage.
Status: not-yet-licensed; do not scrape or ingest
Apartment operating fundamentals, effective rents, occupancy, supply and absorption depth by market/submarket.
Status: not-yet-licensed; do not scrape or ingest
Long apartment rent, vacancy, inventory, and market fundamentals history, often back to the 1990s.
Holland Market Research · HMR-40
Powered by TenureOS
http://10.0.0.123:5110/