Is pre-construction pulling buyers away from Toronto resale?

The thesis is that compelling presale pricing is drawing buyers out of the Greater Toronto resale market. We test it three ways: by the volume of new-home sales against resale sales, by pricing 7,043 available presale floor plans in 661 actively selling projects against an equivalent resale home, and by checking whether resale volumes weakened where and when presale became cheaper.

• Volume. New homes were 15.1% of GTA sales in the first eight months of 2026 (7,505 of 49,625), double 2025's record low but 10 points below the long-run norm. The rebound is almost entirely low-rise: new single-family sales rose to 5,550 from 2,081, while new condo sales (1,955) remain near the lowest on record.
• Price. After the 2026 HST relief, the median presale condo asks 1.58× the price of the same unit brand new on the resale market, and only 4% of condo floor plans are cheaper. Houses and towns are close to parity (0.99×), and 51% of low-rise plans are cheaper than resale.
• Diversion. Resale freehold sales did not weaken relative to resale condos once the HST relief began. The estimated shift is −3.3% (standard error 3.0%), smaller than 7 of 10 placebo years. The surge in new low-rise sales looks largely additive.
• Verdict. The thesis holds for houses and towns: presale was already within a few percent of resale at sticker, and the April 2026 HST relief tipped about half of all low-rise plans below it. Even there, resale volumes show no measurable drain. For condos the evidence points the other way: resale is far cheaper, and presale's share of condo sales has collapsed accordingly.

New and existing homes are substitutes. A buyer choosing a two-bedroom near a subway line can sign a purchase agreement on a tower that will register in three years, or close on a finished unit next month. In a market at equilibrium, the two prices are tied together: the presale is a forward contract on a home, and its price should sit near the cost of buying an equivalent finished home today, carried to the delivery date and adjusted for the rent the forward buyer gives up and the risks they bear.^[16][17][19]

When that link breaks, volume moves. If presale is cheap relative to resale, buyers shift into presale and resale volume sags; if presale is expensive, the reverse happens and presale sales collapse. Studies of linked presale and existing-home markets find exactly this price and volume interaction,^[20] and the stock-flow tradition treats new construction as the margin that adjusts when prices diverge from replacement cost.^[21][22]

The thesis makes three testable predictions:

1. H1, volume: pre-construction's share of all sales should be rising, and high by historical standards.
2. H2, price: a meaningful share of presale homes should be priced at or below an equivalent resale home, once tax, closing costs and timing are treated consistently.
3. H3, diversion: resale volume should weaken in the segment, and at the time, that presale becomes competitive.

2026 is a natural test. On April 1, Ontario and Ottawa removed the full 13% HST on new homes for all buyers, up to $130,000 per home, for agreements signed before March 31, 2027.^[5][6] Resale homes never carried HST. If presale pricing were within a tax wedge of resale, the relief should have tipped the balance.

Four datasets are combined. Each answers a different part of the question and each has its own definitions, which matter when they are put side by side.

Footprint. TRREB's headline counts include Orangeville and five south Simcoe municipalities that BILD excludes; they were about 4% of TRREB sales in September 2026. To check that this does not bias the share, we rebuilt resale volume from MLS® records on BILD's exact five-region footprint for 2003 to 2026. Shares computed both ways differ by at most 1.3 percentage points in any year from 2008 to 2025 (Figure 2 data table).

Restatement and revisions. In April 2025 TRREB restated its whole history to include transactions by members of other boards trading in the GTA, raising annual totals by 4% to 6%. We use the restated totals throughout and scale earlier type splits and pre-April 2025 monthly figures by each year's restatement ratio. Altus revises new-home sales too (2023 has moved from 19,252 to 18,126); annual and year-to-date figures use the latest vintage.

Segments. BILD condo apartments are matched to TRREB condo apartments, co-op and co-ownership apartments; BILD single-family is matched to everything else in TRREB (detached, semi, link, freehold and condo townhouses). The match is close but not exact: BILD counts stacked towns as condo apartments, TRREB counts them as condo townhouses.

For most of the past two decades, new homes were about a quarter of everything sold in the GTA. Between 2008 and 2025 the share averaged 24.9%, peaking at 32.9% in 2011 and 31.3% in 2017. Then volume fell away: new-home sales dropped from 46,479 in 2021 to 5,314 in 2025, the weakest year in Altus's records, while resale fell by about half (51%). In 2025 new homes were 7.9% of all sales.

2026 has brought the first real recovery. Through August, Altus counts 7,505 new-home sales against TRREB's 42,120 resales: a share of 15.1%, up from 7.6% in the same months of 2025. Over the trailing twelve months the share is 12.2%. By any measure that is high relative to 2025 and low relative to history (Figure 2).

The aggregate hides two opposite stories. Condo apartments were the engine of the presale market: in 2017, presales were 56.2% of every condo apartment sold in the GTA, and the 2011 to 2025 average was 43.2%. In 2025 that share was 11.1%, and so far in 2026 it is 15.0%. Ground-oriented homes never had a large presale share, averaging 13.3%, but in 2026 it has climbed to 15.2%, the highest since 2020 and above the long-run norm (Figure 3).

The monthly series shows when the low-rise recovery began. New single-family sales averaged about 219 a month from April to August 2025. In April 2026, the first month of the HST relief, they reached 901, and they averaged 821 a month from April to August, 3.7 times the year before. New condo sales barely moved (Figure 4).

On H1, the answer is mixed. Pre-construction's share of the market has doubled, so the direction is right. But the level is well below its historical norm, and the gain is concentrated in one segment. A market where presale was outcompeting resale across the board would not show condo presales at a third of their usual share.

Comparing a presale price list with resale averages is misleading. Presale units are smaller, newer, sold without parking, priced with HST included, and delivered later. Each difference pushes the comparison in a known direction, so we price every available presale floor plan against a constructed resale twin: the same size, bedrooms and bathrooms, in the same location, brand new, without parking or a locker, at September 2026 prices.

Step 1: what the resale market pays for each feature

We estimate a hedonic price model^[23][26] on 8,683 condo apartment sales and 24,329 house and townhouse sales recorded on the MLS® System between 2026-04-19 and 2026-10-19:

D collects bedroom class (studio through three-plus, with dens), bathrooms, parking spaces, locker, and building age in six bands; for houses it also carries dwelling type and a double-garage flag. λ is a fixed effect for each of 146 condo and 387 low-rise neighbourhoods, and μ for each month. Size is the midpoint of the MLS® range. Building age, missing on most condo listings, is filled by matching the street address to Red Bricks' record of when each building was completed, which raises age coverage for condo sales to about 86%.

Step 2: location, from the nearest sales

Neighbourhood boundaries are coarse, so each presale project's location value is taken from the 25 nearest resale sales of the same type: for each sale we strip out what its own features explain (L_j = ln P_j − X_jβ − μ_month) and average what remains with Gaussian distance weights. The resale twin of floor plan i is then

Validation. Fitted on a random 80% of sales and tested on the rest, the condo model's median absolute error is 7.9%; 59% of predictions land within 10% of the actual sale price and 77% within 15%, with no systematic bias (median signed error −0.1%). The low-rise model performs the same (7.9% median error). That is about as accurate as a resale estimate gets without seeing the unit.

Figure 6 matters for the comparison. A condo loses about 9% of its value in its first decade and about 18% by its third, so a fair comparison for a presale condo is a new resale unit, not the average one. Houses barely depreciate because most of their value is land.^[24][25] We use the brand-new twin as the benchmark, which is the comparison most favourable to presale.

Step 3: putting the two prices on the same basis

We then make four adjustments, reported cumulatively.

1. Sticker. Presale asking price against the resale twin. Ontario presale prices include HST net of the standard rebates the builder claims.
2. Closing costs. New homes carry closing adjustments resale homes do not: development charge and levy pass-throughs (we use each project's published cap, or $8,000 where none is published), plus $3,500 for Tarion enrolment, utility meters and connections. Land transfer tax applies to both and nets out.
3. HST relief. For agreements signed from April 1, 2026 to March 31, 2027, eligible buyers recover the full 13% HST on homes up to $1 million, $130,000 between $1 million and $1.5 million, phasing down to $24,000 at $1.85 million.^[5] We back out the pre-tax price from the sticker and apply the schedule. The median saving is $69,059 on a condo and $82,643 on a house. Owner-occupiers qualify if the home is substantially complete by the end of 2031; investors only if it is complete by the end of 2029,^[7] which excludes most pre-construction units. Results for investors are reported separately.
4. Forward parity. A presale buyer pays a deposit now and the balance at occupancy, τ years away, and rents elsewhere in the meantime. The fair price for delivery at τ is the resale twin carried forward at the cost of money r less the net rental yield y. The buyer also bears completion, appraisal and liquidity risk, priced at π per year:

We set r = 4.0%, y = 3.0% for condos and 2.5% for houses, π = 1.5% per year, and δ from each project's published deposit structure (15% for condos and 10% for houses where none is published). The sensitivity of results to these choices is in Table 7.

Across 3,765 available condo floor plans in 308 projects, the median presale asks 1.70× its resale twin at sticker price. Net of the HST relief it is 1.58×. Just 3.9% of floor plans are cheaper than resale after the relief, and 1.5% are more than 10% cheaper. For houses and towns (3,278 plans, 353 projects) the median is 1.05× at sticker and 0.99× net of relief, and 51.3% of plans are below resale (Figures 7 and 8). Across all 7,043 plans, 25.9% can be bought for less than an equivalent resale home once the relief is counted (18.7% at sticker), and 41% of projects have at least one plan that can.

Each adjustment moves the needle by a known amount (Figure 9). Closing costs push presale up by about 1.5%. The HST relief is the largest single factor, cutting the ratio by roughly 8.3% for condos and 6.6% for houses. Forward parity barely moves the median, because most available inventory is in buildings already complete or close to it, and because the cost of carrying a deposit roughly offsets the rent saved by buying later. Under investor rules the condo median is 1.61× and the low-rise median 1.05×.

A second method. To check that the result is not an artefact of the model, we repeat the comparison with simple local comparables: each plan's net price per square foot against the median price per square foot of resale sales in buildings ten years old or newer, within 1.5 km for condos and 3 km for houses, with the same bedroom class. This method is cruder (resale comparables often include parking, which flatters resale), but it needs no regression. It lands in the same place.

The per-square-foot ladder makes the gap concrete. In the City of Toronto, presale condos ask a median $1,325 per square foot, or $1,223 net of HST relief. Resale condos in buildings five years old or newer sold for $864, and the model's parking-free new twin is worth $786. Independent figures agree: Urbanation puts new, unsold GTHA condo asking prices at $1,157 to $1,189 per square foot through 2026 against $830 to $859 for resale in buildings registered within three years.^[9][10][11]

Two patterns stand out. First, the discount, where it exists, is in ground-oriented homes, and mostly in the 905, where builders compete directly with resale subdivisions of similar age and the HST relief is worth the most. Second, the condo units closest to resale are in completed buildings that did not sell out. Those are new homes in name, but economically they are resale: delivered now, with no completion risk, and priced to clear. For houses the pattern runs the other way:pre-construction plans are the cheapest relative to resale (0.95×), many in phases launched since April and priced for the post-relief market. Urbanation reports that bulk investors bought 339 such units in the third quarter at about $763 per square foot, roughly 40% below original prices and in line with resale.^[11]

Table 6 lists the projects where the model finds presale cheapest relative to resale, among projects with at least two available plans and resale comparables within three kilometres. These are model estimates of value, not appraisals or recommendations. Each project's location, finishes, lot, exposure and incentives deserve their own diligence, and asking prices change weekly.

If cheaper presale were pulling buyers away, the evidence should show up in resale volume: freehold resales should weaken relative to condo resales after April 2026, because the relief made new houses competitive while new condos stayed far above resale. New single-family sales added roughly 3,009 units from April to August compared with a year earlier, against 22,591 resale houses and towns. Full diversion would have cut resale low-rise volume by about 13%.

We test this with the ratio of freehold to condo resale sales on BILD's footprint, monthly from January 2015, controlling for month of year, a quadratic trend and the COVID spring of 2020, with a dummy for April 2026 onward and Newey-West standard errors. Using condo resale as the control removes market-wide forces such as interest rates and confidence, which hit both segments.

The estimated effect is −3.3% with a standard error of 3.0%: small, and not distinguishable from zero. To judge whether even that is unusual, we ran the same regression with a fake event in April to September of each earlier year (Figure 17). 7 of the 10 placebo estimates are as large or larger in absolute terms; the median placebo is 4.5%. Nothing about 2026 stands out.

On H3, the answer is no. The new-home rebound in houses and towns appears to be mostly new demand: buyers who were not transacting at all, pulled in by a five- to six-figure tax saving with a deadline, rather than buyers who would otherwise have bought resale. Some diversion surely happened at the margin, but it is too small to see against normal seasonal noise. The result is consistent with research showing that new supply reshapes demand through moving chains rather than one-for-one substitution.^[27][28]

If presale condos cost about half again as much as the equivalent resale unit, why not cut prices? Because most of that price is cost, not margin. Land bought in 2021 and 2022, construction contracts, financing and development charges set a floor. Lenders typically want about 70% of units presold before releasing construction financing,^[15] and selling below cost erodes the equity that loan rests on. When replacement cost exceeds what the market will pay, the textbook outcome is that new construction stops.^[22] That is what is happening: Urbanation counted zero GTHA condo launches in the first quarter of 2026, the first time in more than thirty years.^[9]

The presale model also depended on a buyer who has largely gone. Investors bought about 70% of presales in recent years, and by 2024 more than 80% of mortgaged new-condo investors were cash-flow negative.^[12] A presale deposit used to work as a leveraged call option on appreciation;^[18][29] with resale prices falling, that option is out of the money, and about one in ten units registering in 2025 failed to close.^[13] The 2026 HST relief does little for this buyer: rental units qualify only if complete by the end of 2029. RBC Economics describes the same mismatch: existing condos are often available for less than developers can viably charge for new ones.^[14] CMHC estimated 57 months of pre-construction supply in early 2025, with investor carrying costs up 24% since 2022 against rents up 15%.^[15]

Low-rise is different on every count. The price gap was already small, end users dominate the buyer pool, the relief is worth up to $130,000 on homes up to $1.5 million, and municipal development charge cuts in Toronto, Vaughan, Mississauga and Peel are largest for houses and family-sized units.^[8] A tax cut worth about 8% of the price is enough to close a 6% gap. It cannot close a 72% one.

How high is pre-construction as a share of total sales? 15.1% of GTA sales so far in 2026, 12.2% over the last twelve months, against a 2008 to 2025 average of 24.9%. Twice last year's level, still well short of normal.

Are compelling presale prices pulling buyers away from resale? In houses and towns the prices are now compelling: 51% of available plans undercut an equivalent resale home after the HST relief. But the extra sales look like buyers who were otherwise sitting out, not buyers taken from resale. For condos, the opposite is true: recent resale is about 35% cheaper per square foot than presale, and buyers have moved accordingly. Presale's share of condo transactions is about 65% below its 2011 to 2025 norm.

• For buyers. In condos, resale is the better value in about 96% of cases; the exceptions are mostly completed, unsold units being cleared at resale-level prices. In 905 houses and towns, about half of presale plans signed before March 31, 2027, with completion before 2032, beat resale once the HST relief is counted.
• For builders. The low-rise rebound has a deadline. Sales pulled forward into the relief window may leave a gap in 2027 if the program is not extended.
• For policy. The HST relief worked where the price gap was within reach of a 13% tax cut. Restarting condo presales needs the gap closed from the cost side: development charges, approvals and construction costs, or a resale market that recovers first.

• Asking versus transacted prices. Presale prices are list prices. Urbanation finds non-bulk new condo sales close about 10% to 15% below asking.^[10] Even a 15% cut leaves the median condo ratio near 1.34×.
• Selection. Available floor plans are what is left; the most desirable stacks may have sold. Sold-out plans are excluded because their prices are stale.
• Measurement. MLS® sizes are ranges and builder sizes are marketing areas. Neither captures view, floor, finishes or amenities, which favour new buildings and so understate the value of presale.
• Advertised prices. Some builders now advertise prices that already net out the 2026 relief. Red Bricks flags those projects and we do not adjust them; any it misses would make presale look cheaper than it is. The sticker results are the bound.
• Dwelling type. Low-rise plans are typed from Red Bricks labels. Untyped plans in projects that sell both houses and towns are excluded rather than guessed.
• Plan weighting. Results weight each available floor plan equally; we do not observe how many units of each plan remain.
• Eligibility. HST relief depends on the buyer and the completion date. We report owner-occupier and investor cases; buyers who qualify for neither should use the sticker results.
• Volume definitions. BILD and TRREB count sales at different moments in the transaction and revise differently. We use each source's latest vintage and test the footprint difference, but small definitional gaps remain.
• Diversion test. Six post-relief months is a short window. The test can rule out large effects, not small ones.

Sensitivity of the forward-parity test

Share of floor plans in buildings not yet complete (τ > 0) that pass the forward-parity test, for alternative spreads between the cost of money and the net rental yield (r − y) and alternative risk premiums π.

Parameters and data handling

• Resale sample: MLS® sales with a recorded size range, condo apartments 300 to 3,000 sq ft and $250 to $3,500 per sq ft; houses and towns 700 to 4,500 sq ft. Range midpoints are used; "0 to 499" is set to 425 sq ft.
• Bedroom classes treat a den as half a bedroom (1 + den = 1.5), matching Red Bricks' floor plan coding.
• Location: neighbourhood fixed effects where a neighbourhood has 20 or more sales, otherwise TRREB district, otherwise region. Spatial smoothing over the 25 nearest same-type sales with a Gaussian kernel whose bandwidth is the median neighbour distance (minimum 250 m).
• Presale sample: floor plans marked available with a current price and size; condo plans below $450 per sq ft and low-rise plans below $300 per sq ft are treated as data errors and dropped.
• HST: sticker P = 1.13B − standard rebates (Ontario new housing rebate up to $24,000; federal GST new housing rebate phased out by $450,000), solved for the pre-tax price B. Projects whose list prices already include the enhanced rebate are not adjusted.
• Time to occupancy τ is measured from mid-October 2026 to the middle of the stated occupancy year; completed buildings have τ = 0.

Replication

The pipeline (data pulls, model, adjustments and dataset build) is kept with the site source. Aggregated datasets behind every figure are published with this article; record-level MLS® and Red Bricks data are licensed and are not redistributed.