Narberth Borough Parking:
A Data-Driven Counter-Analysis

Challenging the ACS-based assumptions used to justify a 0.7 spaces/unit parking standard and zero-parking exemptions for affordable and small-unit housing.

⚠ Critical Review of Proposed Zoning Policy

Executive Summary

Adam Krom presented American Community Survey (ACS) data to support a proposed parking ordinance requiring only 0.7 parking spaces per unit, with zero-parking requirements for affordable housing and units under 750 sq ft. This analysis demonstrates that the data is statistically unreliable for policy use, misrepresents actual car ownership, and would create a significant parking deficit by the proposal's own math.

88%
Renters with at least 1 car (ACS data, Krom's own source)
160%
Margin of error for "3-car renter" households
−14.2
Parking space deficit at 198 Elmwood Ave under proposed rules
95%
Local survey respondents who own at least 1 car
Bottom line: The proposed 0.7 spaces/unit standard, combined with zero-parking exemptions, would fail to meet parking demand by the proponent's own data — resulting in at least 14 unaccounted-for vehicles at a 40-unit development like 198 Elmwood Ave.

1. The Data as Presented by Krom

Krom cited the U.S. Census Bureau's American Community Survey (ACS) 5-Year Estimates (Table B25044) for Narberth Borough (ZIP/place code 42526), presented as a cross-tabulation of vehicle availability by tenure. The table below reproduces his figures exactly.

Krom's Cross-Tab (ACS 5-Year Estimates, 2024 Release)

Category Owners Renters Total % of Total
Total Households1,1078942,001100%
With 0 Vehicles098985%
With 1 Vehicle41358299550%
With 2 Vehicles56019275238%
With 3 Vehicles8710975%
With 4 Vehicles4712593%
With 5+ Vehicles0000%
What Krom's framing emphasizes: "76% of renters have 1 or 0 cars" — used to argue low parking demand and justify the 0.7 standard and exemptions.
What that framing obscures: Only 11% of renters have zero cars. The remaining 65% who have exactly 1 car still need a parking space. Grouping them together is a statistical sleight of hand.

2. The Margin of Error Problem

ACS estimates for small geographies like Narberth Borough (~4,400 residents) carry substantial margins of error. Narberth is a small place — the ACS samples a fraction of its households. This is not a critique of the Census Bureau; it is a fundamental limitation the Bureau itself acknowledges. Using these estimates as the basis for binding zoning policy is statistically inappropriate.

Full Margin of Error Breakdown (Renter-Occupied Households)

Category Estimate Margin of Error (±) MOE as % of Estimate Reliability
Total Households 2,001 ±159 7.95% Acceptable
Owner Occupied 1,107 ±149 13.46% Marginal
Owner – 1 Vehicle 413 ±109 26.39% Unreliable
Owner – 2 Vehicles 560 ±137 24.46% Unreliable
Renter Occupied 894 ±193 21.59% Unreliable
Renter – No Vehicle 98 ±80 81.63% Statistically Invalid
Renter – 1 Vehicle 582 ±149 25.60% Unreliable
Renter – 2 Vehicles 192 ±132 68.75% Statistically Invalid
Renter – 3 Vehicles 10 ±16 160.00% Statistically Invalid
Renter – 4 Vehicles 12 ±18 150.00% Statistically Invalid
When a margin of error exceeds 30%, the Census Bureau's own guidelines flag the estimate as unreliable for policy use. Six of the ten renter sub-categories exceed this threshold. Three exceed 68%, and two exceed 100% — meaning the true value could plausibly be zero.

Margin of Error by Category

Red line = 30% threshold (Census Bureau reliability benchmark). Values above are flagged as unreliable for policy decisions.

Renter Vehicle Ownership: Estimate vs. Possible Range

Error bars show ±MOE. Notice how "No Vehicle" could range from 18 to 178 households — a 10× swing.

Key implication: The 98-household "no vehicle" estimate could be anywhere from 18 to 178 households. Narberth Borough Council member Cyndi Rickards noted this "does not reflect what our apartment managers report" nor anecdotal observation on surrounding streets. The error bars are simply too large for the precision this policy demands.

3. What the Renter Data Actually Shows

Krom's "76% have 1 or 0 cars" framing groups households that need parking (1 car) with those that don't (0 cars). Correctly disaggregated, the data tells a very different story.

Krom's Framing

"76% of renters have 1 or 0 cars" — conflates car owners with non-owners to minimize perceived demand.

Accurate Framing

88% of renters own at least one car and need at least one parking space. Only 11% own no car.

Side-by-Side: Two Ways to Read the Same Data

Framing Krom's Claim Accurate Reading
Renters with 0 cars Bundled into "76%" 11% (98 of 894)
Renters with 1 car Bundled into "76%" 65% (582 of 894) — these people need parking
Renters with 2+ cars ~24% 24% (214 of 894) — these people need 2+ spaces
Renters who own a car (need parking) Not stated 88% (796 of 894)
Implied parking spaces needed per unit 0.7 (proposed) ≥0.88 (minimum, single-car owners only)
No car (11%)
11%
1 car — need 1 space (65%)
65%
2+ cars — need 2+ spaces (24%)
24%

4. Case Study: 198 Elmwood Ave

Using a hypothetical 40-unit development at 198 Elmwood Avenue as a concrete test case, we can demonstrate that the proposed parking formula fails on its own terms — using Krom's own data.

40
Total units (proposed)
4
Affordable units (exempt from parking)
6–10+
Under-750 sq ft units (exempt)
35.2
Parking spaces needed (minimum, by Krom's own data)

Parking Supply vs. Demand at 198 Elmwood Ave

Scenario Exempt Units Spaces Required (0.7×) Spaces Needed (88% ownership) Deficit
Krom's Proposed Standard
4 affordable + 6 small-unit exempt		 10 21 spaces
(0.7 × 30 non-exempt units)		 35.2 spaces
(88% × 40 units)		 −14.2 spaces
Maximum Possible Under Proposal
Only 4 affordable units exempt, all others >750 sq ft		 4 25.2 spaces
(0.7 × 36 units)		 35.2 spaces −10 spaces
Industry Standard (1.0/unit) 0 40 spaces 35.2 spaces +4.8 buffer
Visual: Parking Gap at 198 Elmwood (Krom's Proposed Standard)
Spaces Needed (35.2 minimum)
Spaces Provided (21)
Gap (14.2 unaccounted vehicles)
This is a hard stop. Even using Krom's own ACS data, assuming every single renter has exactly one car (minimizing demand), the proposed 0.7 standard with zero-parking exemptions would leave at minimum 14 vehicles without a dedicated space at a 40-unit building. Because some renters have 2+ cars, the actual shortfall is likely higher.

5. The Transit-Oriented Development Argument

A likely response to parking concerns is that Narberth qualifies as a Transit-Oriented Development (TOD) zone, which often allows reduced parking. This argument fails on several grounds.

What is TOD, Really?

There is no single, legally binding federal or Pennsylvania definition of Transit-Oriented Development. Definitions vary widely across municipalities, MPOs, and state agencies. The most commonly cited standard:

  • Gold standard: Within ½ mile (~10-minute walk) of a high-frequency rail or bus rapid transit station
  • Commonwealth of PA precedent: Has used ½ mile from rail or bus stop in some contexts
  • Problem: Under the broadest interpretation (any bus stop), nearly the entire Borough qualifies — making TOD designation meaningless as a planning distinction

TOD Definition: What Qualifies?

Narberth train station is a SEPTA Regional Rail stop (Paoli/Thorndale line) — not a high-frequency urban rail system.

Car Ownership vs. Transit Proximity

High car ownership persists in suburban rail communities. Narberth is a commuter suburb, not an urban transit hub.

TOD Criteria Krom's Implied Position Reality
Transit type Narberth SEPTA station = TOD qualifier Commuter rail, 1-2 trains/hr peak — not high-frequency transit
Geographic scope Station area = reduced parking zone Under broadest PA interpretation, all of Narberth is "TOD" — making it useless as a standard
Car ownership behavior Transit access reduces car ownership ACS data shows 88% of Narberth renters still own cars despite being near transit
Comparable standards 0.7 spaces/unit appropriate for TOD 1.0 spaces/unit is the typical suburban minimum; industry standards for suburban multifamily range from 1.0–1.2
The TOD argument is a door, not a wall. Invoking TOD as justification for minimal parking opens the door to applying the same standard across the entire Borough — since all of Narberth is within ½ mile of bus stops. This is precisely the risk of using an undefined standard in zoning code.

6. Local Survey Data: What Residents Report

A neighborhood survey of Narberth Borough residents — conducted independently — collected 55 responses from households across the Borough. While not a statistically representative sample, the survey provides hyperlocal ground-truth data that directly contradicts the ACS estimates being used to justify the proposed parking standard.

Note: Cyndi Rickards (Borough Council) independently noted that the ACS data "does not reflect what our apartment managers report" and does not match anecdotal observation. This survey corroborates that concern.

Here's what local sites have to say about parking at their facilities:
  • The Elm — 114 Forrest Ave, Narberth, PA 19072: "I can assure you that every tenant in the property is using their parking space which is included in rent." & "Regarding street parking permit requests, several of our residents do have two vehicles."
  • 100 Forrest Ave: "We will now be 100% full for parking because we are about to lease our last unit."
  • 203 Haverford Ave: "We still have a couple spaces left but they will be filled up once we lease our last remaining units, we have three left."
  • 36 North Narberth Ave: "We have about 11 remaining but upon fully leasing the building they will be filled up. So it's just a temporary situation because we only have one unit leased at the moment."
55
Survey responses (deduplicated by household)
95%
Respondents with at least 1 car
1.87
Average cars per responding household
5%
Respondents with 0 cars (note: survey skews to homeowners)

Survey: Car Ownership Distribution

55 deduplicated Narberth household responses.

ACS Estimate vs. Survey Results

% of renter/resident households with 0 cars vs. 1+ cars.

Density Preference Rankings

Respondents ranked four development scenarios from 1 (most preferred) to 4 (least preferred). The results are unambiguous: Single Family (2–4 homes) is overwhelmingly the most preferred, and Max Density (30–40 units) is overwhelmingly the least preferred.

How to read this: A lower average rank score = more preferred. Score of 1.0 would mean every respondent ranked it #1. Score of 4.0 would mean every respondent ranked it #4 (least preferred).

Average Rank Score by Density Option

Lower score = more preferred. Single Family averages 1.25 — near-unanimous first choice. Max Density averages 3.93 — near-unanimous last choice.

How Many Times Each Option Was Ranked #1 (Most Preferred)

Out of 55 respondents. Single Family received 45 first-place votes. Max Density received just 2.

Full Rank Distribution — Max Density (30–40 units)

How respondents distributed their rankings for the highest-density option.

Full Rank Distribution — Single Family (2–4 homes)

How respondents distributed their rankings for the lowest-density option.

Summary Table: Average Rank by Density Option

Density Option Avg Rank Score # Ranked 1st # Ranked 2nd # Ranked 3rd # Ranked 4th (Least) Community Verdict

Individual Responses (Anonymized)

Caveat: This survey is not a random sample. Respondents self-selected and may skew toward car-owning households. However, its directional finding — significantly higher car ownership than the ACS estimates — aligns with what Borough officials and apartment managers report, and suggests the ACS figures undercount vehicle ownership in Narberth.

7. National & Regional Benchmarks

How does the proposed 0.7 standard compare to actual parking requirements in comparable suburban Philadelphia communities and national affordable housing standards?

Community / Standard Parking Requirement Notes
Krom's proposal — Narberth0.7 spaces/unit (non-exempt)With zero-parking for affordable & <750 sq ft units
Narberth Borough (current)1.0–1.5 spaces/unitExisting zoning baseline
ITE Parking Generation Manual (suburban multifamily)~1.0–1.2 spaces/unitIndustry standard for suburban apartment
HUD minimum (affordable housing)1.0 spaces/unitTypical FHA/HUD underwriting assumption
Typical TOD reduction (urban, high-frequency rail)0.5–0.8 spaces/unitApplies to true urban TOD, not suburban commuter rail
Narberth survey average1.87 cars/householdLocal ground truth, all respondents

Parking Standards Comparison

Krom's 0.7 proposal vs. comparable community and industry standards.

Spaces Needed at 198 Elmwood Under Each Standard

For a 40-unit building, how many spaces different standards would require.

8. Summary of Identified Flaws

Flaw 1: Statistically unreliable source data. Six of ten renter sub-categories in the ACS table have margins of error exceeding 30%, and three exceed 68%. The Census Bureau's own guidelines flag estimates above 30% MOE as unreliable for policy use. Three estimates have MOEs exceeding their point estimate — they are statistically indistinguishable from zero.
Flaw 2: Misleading data framing. The "76% have 1 or 0 cars" claim combines car owners with non-owners to minimize perceived demand. By Krom's own data, 88% of renters own a car and need parking. This is the operative figure for a parking ordinance.
Flaw 3: The math fails by its own data. At 198 Elmwood Ave (40 units, 4 affordable, 6+ under 750 sq ft), the 0.7 standard produces 21 required spaces. Krom's own 88% car-ownership figure requires at least 35.2 spaces — a minimum deficit of 14 vehicles with nowhere to park.
Flaw 4: Zero-parking exemptions compound the problem. Affordable units and small units are, if anything, more likely to have single-car households (one car, not zero). Exempting them entirely from parking requirements ignores that 88% of renters — including affordable renters — own cars.
Flaw 5: TOD designation is overbroad and undefined. Using the broad "within ½ mile of transit" definition, all of Narberth Borough qualifies as TOD — making the designation meaningless as a parking-reduction justification. True TOD standards apply to high-frequency urban transit, not suburban commuter rail with hourly headways.
Flaw 6: Local data contradicts the ACS estimates. Borough apartment managers, Council members' anecdotal observation, and the neighborhood survey all report higher car ownership than the ACS estimates suggest. The hyperlocal data — not a broad regional survey — should inform Narberth-specific zoning.

9. Recommendations

1. Commission a hyperlocal parking study. ACS estimates for small geographies are unsuitable for binding zoning policy. A professional parking utilization study of existing Narberth multifamily properties would provide actionable, reliable data.
2. Reject zero-parking exemptions. No evidence supports the claim that affordable or sub-750-sq-ft renters have materially lower car ownership rates in Narberth. All units should have a minimum parking requirement.
3. Set a minimum of 1.0 space per unit. This aligns with HUD standards, ITE suburban guidelines, and is the floor cited by affordable housing developers. Krom's 0.7 standard falls below every comparable benchmark except true urban TOD.
4. Define TOD explicitly before applying it as a standard. Any ordinance using TOD as a basis for parking reduction must define the geographic scope, transit frequency threshold, and applicable route types — or it will be applied borough-wide with no limiting principle.
5. Engage apartment managers directly. Narberth's existing apartment managers have firsthand knowledge of tenant vehicle ownership. Their operational data — which Borough Council already notes contradicts the ACS — should be incorporated into any parking study.

10. What Narberth's Own Planning Documents Say

Narberth Borough has invested significantly in its own long-range planning. Key documents from the Borough's own planning process contradict the premise that minimal parking is appropriate — and establish that the community itself prioritized parking and character preservation.

Narberth 2040 Comprehensive Plan (Adopted April 2019)

Adopted after an "extensive public engagement process," the 2040 Plan serves as the Borough's binding long-term guide for growth and development. The Plan explicitly aims to "maintain the Borough's character" and ensure "growth and change occur in ways that strengthen its walkable downtown."

The 2040 Plan was developed through direct resident input — the same constituency that the ACS estimates (with their massive error bars) fail to accurately represent. Any zoning amendment that contradicts the community priorities established in that Plan should require a new, equivalent level of public engagement before adoption.

Key tension: Reducing parking below demand doesn't strengthen a walkable downtown — it pushes parking conflicts onto surrounding residential streets, undermining the character the Plan sought to protect.

Source: narberthpa.gov/government/projects/narberth_2040_comprehensive_plan

Zoning Recodification Process (2025–2026)

The Planning Commission has been working since mid-2025 on a "policy-neutral" recodification of the Zoning Code — explicitly described as intended to simplify structure, not change policy. Borough Council's directive to the Planning Commission in August 2025 directed them to address "future land uses and housing" — not to unilaterally reduce parking minimums using contested ACS data.

The Planning Commission recommended the recodification to Borough Council in April 2026 without the changes to civic institutional buildings — suggesting the Commission is exercising restraint on contested amendments. A parking standard change of this magnitude deserves the same level of scrutiny.

A Zoning Open House was held February 26, 2026, to gather community feedback on 5B and 4A zoning discussions. The results of that public process — not a single data presentation — should inform parking policy.

Source: narberthpa.gov/departments/building_and_zoning/ordinances.php

230 Haverford Ave — A Precedent Already Set

The Borough recently granted Preliminary/Final Land Development approval for a redevelopment at 230 Haverford Avenue. That application required full parking and land development review, including a dedicated Parking Lot Parcel (258 Haverford Avenue) as part of the approval package. The developer was required to include an Emergency Vehicle Maneuvering Diagram, Trip Generation Comparison, and full engineering review — demonstrating that the Borough's existing process already demands rigorous parking analysis case-by-case.

The question is: why would the Borough now adopt a blanket low parking standard for future projects that bypasses the rigorous review that 230 Haverford required?

Source: narberthpa.gov/government/projects/230_haverford_ave_proposed_development.php
Bottom line: Narberth's own planning documents prioritize community character, rigorous review, and public input. Adopting a parking standard based on statistically unreliable ACS data — without a hyperlocal study and full public process — is inconsistent with the standards the Borough has applied to every other major development decision.

11. Why Small-Town ACS Data Can't Drive Zoning Policy

This isn't a criticism of the Census Bureau. The ACS is a well-designed survey — but it is explicitly designed for large-geography analysis, not small-municipality policy decisions. Using it for Narberth (population ~4,400) is like using a bathroom scale to weigh a package to the nearest gram.

ACS Sample Size: What's Actually Being Measured

The 5-year ACS for Narberth surveys a fraction of households. With ~2,001 estimated households and typical ACS sampling rates for small places (~15–20%), the underlying sample is roughly 300–400 households — spread across 5 years.

Census Bureau's Own MOE Guidance

The Census Bureau uses Coefficient of Variation (CV) to flag unreliable estimates. A CV above 40% means "use with extreme caution." Most Narberth renter sub-categories exceed this.

The "Small N" Problem

Consider what the ACS data actually represents for renter sub-categories:

The Census Bureau explicitly states: "The ACS is designed to provide reliable data for areas with populations of 65,000 or more for 1-year estimates." For smaller areas, 5-year estimates are used — but even these carry substantial uncertainty for sub-group analyses. Narberth's population of ~4,400 is nearly 15× below the threshold where 1-year ACS data would even be published. Using vehicle-availability sub-categories of 894 renter households as the basis for a binding zoning standard falls well outside the ACS's intended policy application.
What the ACS is designed for: Identifying broad regional trends, allocating federal funding, tracking decadal change. What it is not designed for: determining the number of parking spaces required at a specific 40-unit building in a 0.5-square-mile borough.

12. The Austin & DC Comparison: Why It Doesn't Apply

TOD parking reductions are commonly cited using examples from Austin, TX and suburban Washington, DC. These are instructive examples — but they actively undermine, rather than support, the case for applying TOD standards to Narberth. Here's why.

The core problem with these comparisons: Austin's TOD zones sit along light rail and bus rapid transit corridors with 10–15 minute headways. DC Metro's suburban stations serve trains every 4–6 minutes at peak. Narberth's SEPTA Regional Rail service runs approximately every 30–60 minutes — a fundamentally different transit product.

Transit Service Comparison: Narberth vs. Cited TOD Examples

Location / Line Transit Type Peak Headway Off-Peak Headway Parking Standard Used Appropriate for Narberth?
Washington DC Metro
Red/Blue/Orange/Silver/Green/Yellow lines		 Heavy Rail (subway) 4–6 min 10–12 min 0.0–0.5 spaces/unit (urban stations) ❌ No — 5–10× more frequent
DC Metro Suburban Stations
Vienna, Franconia, Shady Grove		 Heavy Rail 6–8 min 12–15 min 0.5–1.0 spaces/unit (with shared parking) ❌ No — still 4–8× more frequent + dedicated park-and-ride
Austin, TX — MetroRail
Red Line commuter rail		 Commuter Rail 30 min (peak only) No off-peak service 0.5–1.0 near stations (Austin's own reductions) ⚠ Similar frequency — but Austin TOD zones are high-density mixed-use, not 0.5 sq mile borough
Austin, TX — MetroRapid
Bus Rapid Transit lines		 BRT 10–15 min 15–20 min 0.3–0.7 near major stops ❌ No — BRT, not commuter rail; dramatically different ridership behavior
SEPTA Paoli/Thorndale Line
Narberth Station		 Commuter Rail (diesel) ~30 min (rush) 60 min (off-peak) Proposed: 0.7/unit ✓ This is the actual baseline

Small Suburban TOD Communities: What They Actually Require

The most honest comparison isn't Austin or DC — it's small, walkable suburbs similar to Narberth that are also near commuter rail. Here's what comparable communities actually require:

Community Population Transit Type Parking Requirement (multifamily) Notes
Haddonfield, NJ ~11,500 PATCO High-Speed Line (every 10–15 min) 1.0–1.5 spaces/unit Higher frequency than SEPTA Regional Rail; still requires 1.0 minimum
Media, PA ~5,700 SEPTA Media/Wawa trolley + R3 rail 1.0 spaces/unit minimum Downtown TOD overlay; 1.0 maintained despite walkable core
Jenkintown, PA ~4,400 SEPTA Regional Rail (similar to Narberth) 1.0–1.25 spaces/unit Comparable size and transit access to Narberth; no 0.7 standard
Collingswood, NJ ~14,000 PATCO High-Speed Line 1.0 spaces/unit (1.5 recommended) Active TOD planning; still requires 1.0 minimum
Ardmore, PA ~13,000 SEPTA Regional Rail + Route 100 trolley 1.0 minimum; structured parking required for large projects More transit access than Narberth; still holds 1.0 minimum
Narberth, PA (proposed) ~4,400 SEPTA Regional Rail only (30–60 min headways) 0.7 (with zero-parking exemptions) Below every comparable community; weakest transit access of the group

Peak Train Frequency: Narberth vs. TOD Comparables

Lower minutes = better service. Narberth has the worst transit frequency of any community whose parking standards are being cited as justification.

Parking Standard vs. Transit Frequency

There is a clear relationship: as transit gets better (lower headway), parking requirements decrease. Narberth's proposed 0.7 standard is misaligned with its actual transit service level.

The Austin/DC comparison is a category error. Citing Washington Metro or Austin BRT as justification for reducing parking in Narberth is like citing Manhattan parking minimums (0 spaces required) to justify eliminating parking in Malvern. Transit frequency, urban density, and ridership behavior are fundamentally different at each scale. Narberth's SEPTA service — one train roughly every 30–60 minutes — places it firmly in the commuter rail suburb category, not the urban transit hub category.

The DVRPC TOD Scoring Framework (Our Own Regional Planner)

The Delaware Valley Regional Planning Commission (DVRPC) — the regional planning body that covers Greater Philadelphia — uses a multi-factor scoring system to evaluate actual TOD readiness. The criteria include transit service quality, job access by transit, travel time competitiveness, residential intensity, car ownership rates, non-car commuters, and Walk Score.

Under DVRPC's Transit Connectivity Index (TCI) framework, a location's TOD score reflects "the number of transit trips accessible by walking." For a SEPTA Regional Rail station with 30–60 minute headways, the TCI score would score in the lowest tier (1–2 out of 4) — far below what DVRPC considers supportive of TOD parking reductions.

DVRPC also scores stations on Job Access (jobs reachable in 30-minute transit ride) and Non-Car Commuter rate. Narberth's suburban bedroom-community character means most residents commute outward by car, not inward by rail — the opposite of the TOD profile that justifies parking reductions.

Source: DVRPC — Evaluating TOD Opportunities in Greater Philadelphia (dvrpc.org/webmaps/tod/) Federal Transit Administration — Transit-Oriented Development Program (transit.dot.gov/TOD)
A note on Austin specifically: Austin's parking reductions apply along corridors served by the MetroRapid BRT (10–15 minute headways) and within ¼ mile of major activity centers — not commuter rail stops. Austin's own land code revision faced significant community pushback precisely because it applied urban standards to suburban contexts. The lesson from Austin is the opposite of what TOD proponents suggest: getting the geographic and frequency thresholds right matters enormously.

Sources & Methodology

Prepared May 2026 · Narberth Borough, PA 19072 · This document is intended for public comment and civic deliberation purposes.