In Ontario, as elsewhere, the systematic application of land use standards began after the Second World War. These standards set requirements for key features of urban form: road widths, residential lot sizes, lot frontages and depths, and water and sewer services. These development standards were part of a larger project of separating residential areas from commercial and industrial areas, segregating different housing types, and disconnecting local streets from arterial roads and expressways (Ben-Joseph 2005; Southworth & Ben-Joseph 1997; Krieger 2005). More generous allocation of land to public facilities necessarily reduces the proportion of the land base given over to private residential and commercial uses, and therefore reduces gross and developable area density.
Literature review
Standards for the allocation of land for parks, schoolyards, roads, and environmental protection are believed to have increased over time in the Toronto metropolitan region (IBI Group 1993; Blais 2000:37; CMHC 1996). Evidence for this claim has typically been based on empirical study of the physical landscape and analysis of plans of subdivision. In a 2000 study that compared five existing 2km-by-2km segments of the GTA, Wright found that "as development has moved from the urban core to the suburbs, there has been ... a continuous increase in the amount of land consumed per hectare of residential area" (2000:96).
Ontario's Planning Act permits municipalities to require as a condition of development conveyance of either a maximum of 5% of residential land area (s. 42(1)) or a maximum of one hectare per 300 dwellings (s. 42(3)) to a municipality for parks or other recreational purposes. A review of Toronto-area official plans found that parkland is commonly specified on a per-1,000-resident formula.
Standards for parks and other public uses have been in place since the passage of Ontario's first modern Planning Act in 1946. Systematically applied standards did not play such a role in shaping prewar urban development. Examination of the text of earlier versions of the Planning Act found that a parkland conveyance standard of one acre per 120 dwellings predates 1960 at the provincial and municipal level.14 The 1959 draft plan for the Metropolitan Toronto Planning Area states that "in some municipalities, statutory park dedications (amounting to 5% of subdivided land) are applied." (The plan further notes that a dedication in proportion to subdivision land area is a poor policy tool, as it is intended to produce 21/2 acres [about one hectare] of parkland per 1,000 residents in a single-detached housing development.15) Arguments in favour of reducing standards for public uses have been made by the property development and house-building industries (Hemson 2003b:13). Since the 1970s, there have been several attempts at reform of development standards in Ontario, although the motivation had less to do with increasing density than with lowering the capital costs of public infrastructure (MHO 1976, 1982; MMAH 1995a).
In a comparison of neotraditional and conventional subdivisions in Markham, Gordon and Vipond (2005:41-54) found that while the neotraditional plans exhibited much higher developable area densities, they did not on average contain more park and schoolyard land as a proportion of total subdivision land area than conventional subdivisions, meaning that the per-capita public land area is lower in neotraditional plans.
Research questions
1. Is the proportion of developable land allocated to public facilities higher the more recently a study area was developed?
2. Do more recently developed areas have more park and schoolyard land on a per-person or per-dwelling basis?
Findings
The balance of public and private land uses
Contrary to expectations, the proportion of the gross land area allocated to private uses (residential and employment parcel area) in the 16 study areas varies little, ranging from 51% in the Peanut to 64% in Milton, with no trend by era of initial development. (See Fig. 17.)
Fig. 17: Private property as % of gross land area
The proportions of public and private property vary more in terms of developable land area, but again, there is no pattern by era of initial development. (See Figs. 18 and 19.) If the study areas with the lowest and highest values -- the Peanut, which has a high proportion due to its "tower-in-the-park" design, and Cachet, which contains a large natural heritage system and narrower-than-average rights-of-way -- are excluded, the variation is only about 10%, and this variation occurs between two pre-1960 cases (Riverdale and Leaside). At first glance, this appears to fit Gordon and Vipond's (2005) finding that public land coverage varied little between neotraditional and conventional subdivisions in Markham. Aggregating all public uses masks significant variation in the proportions accounted for by subcategories, however.
Parks and open space as a proportion of gross land area
Looking at parks in isolation, there is a clear break between pre- and post-1960 development. In the pre-1960 study areas, parks cover less than 5% of developable area. The proportion is considerably higher in most later-era study areas, although there is no discernable pattern by era group. (See Fig. 20.)
A further test was conducted to determine whether this finding is due to the presence of environmentally protected lands that function as public open space. Parks, hazard lands, and environmentally protected lands were combined into a generic "open space" category. As Fig. 21 shows, this also yielded no clear pattern by era group, because of large variations within each group.
Fig. 18: Private property as % of developable land area
Fig. 19: Public facilities as % of developable land area
Includes parks, schoolyards, places of worship, cemeteries, and rights-of-way.
Fig. 20: Parks as % of developable land area
Fig. 21: Open space as % of gross land area
Open space includes parks, hazard lands, and environmentally protected lands and excludes highway and rail corridors.
Fig. 22: Schoolyard land as % of developable land area
Excludes Richmond Hill.
Schoolyards as a proportion of developable land area
The proportion of developable land area used for schoolyards also varies greatly, with no pattern by era of development. Some study areas have high or low values due to anomalous situations. The Milton study area, for example, features a school for the blind and developmentally disabled that caters to a non-local population. The Richmond Hill study area contains no public or Catholic schools, although several lie just beyond its borders. (For this reason, the Richmond Hill study area is excluded from Figs. 22 and 25.) This finding may indicate a trend towards larger service areas for schools or may be an artifact of the boundary-drawing process.
Milton and Richmond Hill aside, there is on average less schoolyard land as a proportion of developable land area in the post-1980 study areas than in previous era groups. This finding may be due to shared-facilities policies for parks and schoolyards. Although such policies were not explored in this project, it appears that parks and schoolyards were co-located in all post-1960 study areas.
Fig. 23: Rights-of-way as % of developable land area
Rights-of-way as a proportion of developable land area
Land areas for rights-of-way range from 18% in Bronte to 34.9% in Leaside. (See Fig. 23.) The Bronte value is low because of the presence of substantial employment and utility land areas that contain few local roads. The next lowest value is for Milton, at 21.4%, probably owing to the configuration of the street and block network, as the superblock developments common in the 1960s and 1970s have fewer roads and therefore less road coverage. The presence of highway access ramps and service roads means that values for Richmond Hill and Vaughan are higher than would otherwise be the case. Values for the pre-1960 study areas vary more than most areas built subsequently, suggesting the convergence on common standards for street widths and street network configuration.
Land area for parks in proportion to population and dwellings
Although there appears to be only a weak relationship between era of development and the amount of land allocated to public facilities, the relationship between the era of development and the land area of parks and schools per capita and per dwelling unit is much stronger. (Land area per capita for rights-of-way was not determined, as the data do not distinguish roads serving residential neighbourhoods from those serving employment lands.)
Of the post-1960 study areas, all but Richmond Hill exceed the 5% standard and all exceed the one-hectare-per-300-dwellings standard. With the exception of Bronte -- which has an anomalously high value because of its small population and large employment zone -- parkland area per capita and per dwelling unit is, in general, higher the more recently a study area was built out. (See Fig. 24.) Higher parkland per capita in more recent cases results in no discernable increase in the proportion of developable land area devoted to parks, because these areas generally have lower population densities. For example, the Mississauga Valleys and Markham Northeast cases have similar gross and residential lot areas, and public facilities as a percentage of developable land area. But since Mississauga Valleys contains 70% more residents than Markham Northeast, it has half the amount of parkland per capita.
Land area for schoolyards in proportion to population and dwellings
Excluding Richmond Hill (which has no schools) and Milton (which contains a special-needs school serving a regional rather than neighbourhood clientele), schoolyard area shows no clear pattern on a per-dwelling-unit or per-resident basis. (See Fig. 25.) The narrower range of values in the post-1980 group suggests that different jurisdictions' standards for schools have converged over time. Land area per dwelling for schoolyards and parks is greater in the more recent study areas, suggesting that the potential for dual-use facilities to reduce the overall amount of land set aside for these uses has not been realized, even though schoolyards tend to be located next to parks in all post-1960 study areas.
The impact of public facilities on gross and developable area density
Fig. 26 shows that residential parcel area as a proportion of gross land area varies significantly across the 16 study areas, although without a clear pattern by era of development. This is due to the presence of major employment lands in Bronte, Richmond Hill, and Vaughan, and protected environmental areas in Glen Abbey. If these are excluded, the residential component ranges from 45% to 55% of the gross land base in the other 13 cases.
What impact then does the proportion of non-residential land have on gross and developable area densities? In a given district, land use distribution is a zero-sum game. Increasing the size of one component necessarily reduces the others. If the net residential parcel area is reduced by raising standards for public facilities or environmental protection, the net residential density must increase for the same gross density to be achieved.
Fig. 24: Parkland area pro rata
A. Parkland in hectares per 1,000 residents (excluding Bronte)
B. Parkland in hectares per 300 dwelling units (excluding Bronte)
Fig. 25: Schoolyard area pro rata
A. Schoolyard area in hectares per 1,000 residents (excluding Milton and Richmond Hill)
B. Schoolyard area in hectares per 300 dwelling units (excluding Milton and Richmond Hill)
Fig. 26: Residential parcel area as % of gross land area
Summary of findings
1. Is the proportion of developable land allocated to public facilities higher the more recently a study area was developed?
No. The proportions of both gross and developable land area accounted for by public facilities and private property vary little across the 16 cases. This analysis does not support the contention that, in aggregate, increasingly generous standards for parks, schools, and roads have depressed gross and developable area density.
However, despite the overall consistency in the proportion of public versus private land, disaggregating the public use categories reveals that the proportions of developable land area accounted for by parks, schools, and roads vary considerably both within and between era groups.
2. Do more recently developed areas have more park and schoolyard land on a per-person or per-dwelling basis?
Yes. In general, the more recently a study area was planned and built, the more parkland area there is per resident and per household, suggesting that parkland area allocations have increased over time. The same is not true of schools. Schoolyard area per capita and per dwelling vary considerably within and between era groups. The variation within each era group decreases with each successive group, however, indicating that standards governing schoolyard size may have become more uniform over time.
Implications for policy
This analysis provides no conclusive evidence of rising development standards. Despite an observed correlation between the era of initial development and the amount of parkland per capita, this finding does not imply a causal relationship. At the same time, the analysis does not disprove the thesis that planning regulations have exerted upward pressure on the amount of land set aside for public uses over time. To make a conclusive causal link, attributes of the built environment must be comprehensively compared to the standards and policies in effect when the study area was originally planned and built. Moreover, as public facilities land is typically specified in proportion to population, such an analysis would have to determine the population initially expected to inhabit the area in question. Retrospective analysis is beyond the scope of this report.
This analysis does not account for regulations that restrict development rights within private parcels. Setback and buffer requirements, for example, may reduce the development capacity of the net private parcel area. It is possible that such regulations have increased over time. This study also could not assess whether natural heritage protection measures at a broader geographic scale have become more generous over time, thereby reducing gross density.
How standards for public uses are specified -- as a percentage of land area, per capita, or per household -- makes a difference. The relationships between land area, population, households, and dwellings change over time. Given the long-term decline in average household size since the 1960s, some of the older study areas had higher populations when they were built than they do today. As a result, already low levels of park and schoolyard land per capita would have been substantially lower when the oldest neighbourhoods were built than they are today. In other neighbourhoods, the reverse is true. Today, high-rise apartment complexes in the Mississauga Valleys, Peanut, and Malvern study areas have become immigrant reception areas with larger-than-average household sizes and may therefore accommodate higher populations now than when they were originally constructed and occupied in the 1960s and 1970s.
Notes
14. The Planning Act, R.S.O. 1960 c. 296 s. 28(5)(a) contains the 5% conveyance standard. The report of the Planning Act Review Committee (1977:119) notes that the 5% figure is derived from a "commonly accepted standard that there should be 21/2 acres of parkland for every 1,000 persons in a residential neighbourhood, and was intended to more or less yield this ratio where neighbourhoods were developed at low, single-family densities. The proportional land yield from the 5 percent dedication is of course much lower with higher densities of development. To overcome this deficiency, another provision of the Act ... allows municipalities to secure parkland dedications at a ratio of one acre for every 120 dwellings."
15. See Metropolitan Toronto Planning Board (1959) 226-30. The 1959 plan proposed a total of 71/2 acres of parks and open space per 1,000 residents region-wide: 21/2 acres of local parks, 31/2 acres of metropolitan parks, and 11/2 acres of undeveloped public open space. This was reiterated in the 1965 plan (Metropolitan Toronto Planning Board (1965) Assumptions s. 23; Objectives s. 15).