How We Estimate Local Costs

Window prices are not the same everywhere, and a single national range is honest but blunt: it's technically correct in Boston and in Fargo while telling neither homeowner much. So on metros where we have the data, we shift the national ranges to the local cost level, and we tie that shift to a public federal index rather than inventing a number. This page explains exactly how, because a cost figure is only as trustworthy as the source behind it.

The method in one line

We start from federal window measure-cost data, add the installation cost of a true replacement, bring it forward to today's dollars, then shift it to the local cost level. Every step is a public dataset.

The four steps, in full

1. Federal measure cost. The U.S. Department of Energy's NREL National Residential Efficiency Measures Database publishes window material cost per square foot at the 10th, 50th and 90th percentile. We use the REMDB 2024 machine-readable release (REMDB_2024.12.23.xlsx), which states its figures in 2023 dollars.
2. × 1.673 for installation. NREL's own retrofit multiplier, which covers labor and removing and disposing of the existing window. That's what makes these installed prices rather than the price of the unit on a pallet.
3. × 1.1193 to reach today. REMDB is not reissued every year, so 2023 dollars drift further from the market each year it sits. We escalate using the U.S. Bureau of Labor Statistics Producer Price Index for window and door manufacturing, averaged across three series, from the 2023 annual average to June 2026. Skipping this step is the single most common reason a cost page reads low.
4. × the local factor. The BEA Regional Price Parity (2024), a U.S. Bureau of Economic Analysis index of how a metro's price level compares to the national average (U.S. = 100). A metro at 108 is about 8% above the national level; a metro at 91 is about 9% below.

The published range for a window style runs from the 10th to the 90th percentile. The low end is a mid-grade vinyl insert from a lean regional installer; the high end is a premium composite or wood-clad unit, a full-frame installation, or a full-service national brand that bundles product, labor and warranty into one price. Both ends are real quotes that homeowners receive.

Why the local difference is real but moderate

Here's the part most cost pages won't tell you. The BEA index breaks into parts, and they don't move together:

• The window itself barely varies by metro. BEA's price parity for goods is close to the national average almost everywhere (Boston 99.7, Fargo 95.7), because windows are manufactured and shipped nationally.
• Most of the "expensive city" gap is housing rent (Boston 148, Fargo 81), which has nothing to do with replacing your windows, so we don't let it dominate the estimate.
• What legitimately moves the price is local install labor and overhead (Boston 104, Fargo 91). That's the real driver of a metro cost difference.

The upshot: a defensible metro adjustment for windows lands around plus or minus 8–10%, not the 40% a cost-of-living headline implies. We'd rather show you a smaller, honest adjustment than a dramatic one we can't stand behind.

The assumptions behind every number

These are the judgement calls the method rests on. They are stated here rather than buried, because a reader who disagrees with one of them can work out exactly how much it would move the figure.

Assumed opening size

REMDB prices windows per square foot, so turning that into a price per opening requires assuming a size. These are standard replacement openings, not measured averages of any particular market. A larger window costs proportionally more, so if your openings run bigger than these, expect to sit above the published range:

• double hung: 15 sq ft
• casement: 12 sq ft
• sliding: 15 sq ft
• picture: 20 sq ft
• awning: 8 sq ft

Frame material multipliers

Vinyl is priced directly from REMDB. Other frame materials are derived from it using market multipliers. These come from 2026 industry cost surveys rather than a federal dataset, and are the softest step in the chain:

• aluminum: 0% to 15% above vinyl
• composite: 25% to 30% above vinyl
• fiberglass: 28% to 45% above vinyl
• wood (clad): 60% to 130% above vinyl

Products REMDB cannot price

Five products are not derived from REMDB at all, because its per-square-foot model does not survive a sanity check on them. Structural work on bays and bows does not scale with glass area, and the door figures produce results no installer would recognise. These keep a market-anchored range escalated by the same PPI factor, and we would rather say so than imply a federal derivation that isn't there:

• Bay & bow
• Specialty shapes
• Sliding patio door
• Hinged / French patio door
• Entry door

What this is, and what it isn't

• Planning ranges, not quotes. We round to the nearest few dollars and present a range; your home's age, the glass package, and the installer's sales model still move you inside (or past) it. Always get at least three itemized bids.
• Conservative by design. The BEA parity is a consumer-price index, so it under-captures the union-labor premiums that specialized construction indexes (RSMeans, ENR) measure. Real installed prices in high-labor-cost metros can run a bit higher than our adjustment suggests. We'd rather err toward the defensible public number.
• Only where we have data. Metros without a published BEA parity keep the national ranges. We never fabricate a local figure.

Sources

• — U.S. Bureau of Economic Analysis, Regional Price Parities by Metropolitan Area, 2024 (all items, goods, services; via Federal Reserve Economic Data, FRED)
• — NREL National Residential Efficiency Measures Database, REMDB 2024 machine-readable release (REMDB_2024.12.23.xlsx); window and door measures, 2023$ basis
• — U.S. Bureau of Labor Statistics, Producer Price Index by industry: PCU332321332321 (metal window & door mfg); PCU321911321911 (wood window & door mfg); PCU326199326199 (other plastics product mfg). Escalation from the 2023 annual average to June 2026.
• — Frame-material multipliers: 2026 industry replacement-window cost surveys (market data, not a federal source)

Pricing model last revised 2026-08-09. Every figure on this site is generated from it; see our national cost guide for the resulting ranges.