AFFORDABLE HOUSING FEASIBILITY STUDY FINAL. Prepared for Parramatta City Council

AFFORDABLE HOUSING FEASIBILITY STUDY FINAL JUNE 2017 Prepared for Parramatta City Council

SGS Economics and Planning Pty Ltd 2017 This report has been prepared for Parramatta City Council. SGS Economics and Planning has taken all due care in the preparation of this report. However, SGS and its associated consultants are not liable to any person or entity for any damage or loss that has occurred, or may occur, in relation to that person or entity taking or not taking action in respect of any representation, statement, opinion or advice referred to herein. SGS Economics and Planning Pty Ltd ACN 007 437 729 www.sgsep.com.au Offices in Canberra, Hobart, Melbourne, Sydney

TABLE OF CONTENTS 1. INTRODUCTION 3 1.1 Background 3 1.2 Scope of our work 3 1.3 Structure of this report 3 2. RESIDUAL LAND VALUE MODELLING 4 2.1 Residual land value analysis 4 2.2 The Model 5 Key parameters 7 3. DEVELOPMENT SCENARIOS, COSTS AND REVENUES 8 3.1 Development scenarios 8 3.2 Revenue and cost assumptions 9 Gross Residual Value per sqm 9 Construction costs per sqm 10 Infrastructure contributions 11 Value capture policy 11 Special infrastructure contribution 11 ARH contributions 12 Purchase cost of land 12 4. FEASIBILITY RESULTS 13 4.1 Key results 13 Feasibility of handing over ARH at no charge 13 Feasibility of handing over ARH at cost 14 4.2 Additional ARH scenarios 15 5. CONCLUSION 19 LIST OF FIGURES FIGURE 1: CHANGE IN RESIDUAL LAND VALUE WITH UP-ZONING 4 FIGURE 2: SNAPSHOT OF RLV MODEL 6 LIST OF TABLES TABLE 1: UNIVERSAL ASSUMPTIONS AND INPUTS USED BY RLV MODEL 7 TABLE 2: CASE STUDY SITES 8 TABLE 3: DEVELOPMENT YIELDS UNDER EXISTING AND PROPOSED CONTROLS 9 Affordable housing feasibility study 1

TABLE 4: ESTIMATED GRV PER SQM OF GFA 10 TABLE 5: CONSTRUCTION COSTS PER SQM 10 TABLE 6: PURCHASE COST OF LAND 12 TABLE 7: KEY RESULTS OF THE FEASIBILITY ASSESSMENT 10% ARH CONTRIBUTIONS AT NO CHARGE 13 TABLE 8: KEY RESULTS OF THE FEASIBILITY ASSESSMENT 10% ARH CONTRIBUTIONS AT COSTS 14 TABLE 9: KEY RESULTS OF THE FEASIBILITY ASSESSMENT 8% ARH CONTRIBUTION AT COST 15 TABLE 10: KEY RESULTS OF THE FEASIBILITY ASSESSMENT 5% ARH CONTRIBUTIONS AT COSTS 16 TABLE 11: RESULTS OF THE FEASIBILITY ASSESSMENT - 15% ARH CONTRIBUTION AT NO COST ON GOVERNMENT SITE 16 TABLE 12: RESULTS OF THE FEASIBILITY ASSESSMENT 10% OF THE VALUE UPLIFT AS ARH 17 TABLE 13: RESULTS OF THE FEASIBILITY ASSESSMENT 15% OF THE VALUE UPLIFT AS ARH 17 TABLE 14: RESULTS OF THE FEASIBILITY ASSESSMENT 20% OF THE VALUE UPLIFT AS ARH 18 TABLE 15: RESULTS OF THE FEASIBILITY ASSESSMENT 5% OF TOTAL DWELLINGS UNDER CURRENT CONTROLS + 10% OF VALUE UPLIFT AS ARH 18 Affordable housing feasibility study 2

1. INTRODUCTION 1.1 Background City of Parramatta Council (CoP) has released an Affordable Housing Discussion Paper for comment. Stakeholder feedback on the Discussion Paper, together with other research, will be used to assist in the development of an Affordable Housing Policy for the whole local government area (LGA). One of the options considered by this Paper to provide for more affordable housing is to impose a 10% of yield requirement for Affordable Rental Housing (ARH) on private land and a 30% of yield requirement for ARH on government land. It is proposed that these requirements are in addition to the current development contributions, any Special Infrastructure Charges and Value Capture Policy (for the CBD) that may apply to the new development. CoP requires a research report that assesses the economic feasibility impacts of introducing such an affordable housing program in the LGA. 1.2 Scope of our work SGS has been commissioned by CoP to undertake this research project. More specifically, our work is to help CoP understand: General impacts and site specific impacts of imposing a 10% of yield Affordable Rental Housing requirement on private land and a 30% of yield requirement for ARH on government land (in context with the current and proposed other levies that could also apply (Sect 94, Value Capture Policy for the CBD, and the SIC) The model in question is based on the developer handing the given number of affordable rental dwellings over to Council at no charge. This should be the base case for feasibility testing, but we also want to understand how feasibility would be impacted if this was modified to an at-cost charge to Council (or whatever body we identify to own the assets) The impacts are to be modelled for six sites in the LGA. 1.3 Structure of this report The rest of this report has been structured as follows: [Chapter 2] Overview of residual land value analysis and key assumptions utilised [Chapter 3] Development scenarios for six sites modelled [Chapter 4] Key findings of the site-level analysis [Chapter 5] Conclusions Affordable housing feasibility study 3

2. RESIDUAL LAND VALUE MODELLING This chapter first provides an overview of the residual land value (RLV) analysis, which will be used in this report to understand the impact of the proposed ARH contribution on development feasibility. Following this, a number of key assumptions used in the RLV modelling are detailed. 2.1 Residual land value analysis Within a number of parameters set by land economics, a developer will determine their best (highest) price for a candidate site on a residual basis. That is, they will start with the expected gross revenues generated from the project upon completion (gross realisation value GRV) and deduct their margin for profit and risk plus all other development costs (construction, approvals, design, marketing etc) to arrive at a residual amount that can be offered to the land owner. If a piece of land is up-zoned, it follows that the residual land value will increase. This is illustrated in the following diagram. In fact, the marginal increase in the land value is a direct measure of the additional development rights conferred by the planning changes. FIGURE 1: CHANGE IN RESIDUAL LAND VALUE WITH UP-ZONING Gross return $ from sales post-rezoning or approval for higher value use and/or density Developer Margin for Profit and Risk Gross return $ from sales/rent pre-rezoning or approval Developer Margin for Profit and Risk Charges and Taxes Construction, marketing and finance costs Land Value Pre rezoning development values Development costs under current zoning, height and FSR. Charges and Taxes Construction, marketing and finance costs Land Value Post rezoning development values Future development costs after rezoning and changes to height and FSR. Costs include construction, charges and taxes and a developer margin for profit and they increase by the increments shown. Land value increases from the base reflecting additional development rights which in turn reflect accessibility to amenities and infrastructure enabled by the rezoning. This value is not created by the land owner or developer and a share is appropriately captured for the community. Source: SGS Economics and Planning, 2017 A number of observations can be made from this diagram in respect of affordable housing targets on up-zoned land. In the absence of any countervailing measure, the full value of the additional development rights will be capitalised into the market price of the development site. That is, the land trader or incumbent owner will gain 100% of this uplift for no particular value adding effort on their part hence the term windfall gain If the up-zoning is accompanied by an additional development cost say a cash or in-kind inclusionary requirement for affordable housing - this will squeeze the increment in land value. Affordable housing feasibility study 4

The gap between this cost impost from the inclusionary requirement and gross value uplift will represent the limit of what the developer will be prepared to pay for additional density (that is, to purchase the full amount of density to fill out the environmental envelope). If measures are put in place for public capture of 100% of the marginal increase in land value occasioned by the up-zoning, the incumbent land owner will have little incentive to release the land to a bona fide developer; this would reduce the supply of development sites into the market and, possibly, reduce the supply of housing as a consequence. Similarly, if market trading in the lead up to the official up-zoning of the land in question causes a bidding up of its price to the full extent of the residual land value attaching to the additional development rights enabled by the official up-zoning, the scope for value sharing including via the provision of affordable housing, will have been removed, at least for the time that it takes for the incumbent owner to adjust to a lower than speculated value for their land. It is not possible to manage the speculative bidding up of development site prices simply by creating more development capacity on the site. This development capacity must be determined by environmental sustainability and desired built form outcomes. From an economic perspective the ultimate development capacity of an up-zoned site is conceptually fixed. Any development capacity added beyond this envelope implies social and environmental costs, which may well offset the value of affordable housing benefits generated in the process. It is also noteworthy from this discussion that the impact of affordable housing targets on project viability does not occur in a static environment. Land traders and developers alike will behave differently depending on how much notice they have these requirements. Clearly, a price taking developer who has purchased a site and is confronted with an unscheduled additional cost is more likely to resist this impost compared to a developer who has the opportunity to factor the cost of the affordable housing requirement into their feasibility study and tendered land price. 2.2 The Model For this project, SGS has developed a spreadsheet-based model, underpinned by the RLV concept discussed above. The model calculates the residual value of a development after deducting all the development costs from the sales revenues, in the current market. The development costs include construction costs and contingencies, professional fees, developer s profit margin, marketing/financing/holding costs, infrastructure levies, value capture charges (where applicable in the CBD), as well as costs (or lost revenue) associated with providing required % of dwellings as ARH. In summary, this model will identify the residual land value for each of the six sites considered, under the following two development scenarios: A highest and best use scenario under the current planning controls The development outcome (with the upzoning) sought by the Planning Proposal received by CoP. The reason both existing and proposed controls will be tested is to address the assumption that in many cases an ARH contribution may not be feasible where land values have been set by the maximum development potential under the current controls. In these cases, the ARH contribution may represent a cost impost not anticipated by the developer. A development scenario can tolerate the ARH contribution, where the RLV after deducting the costs of providing the required % of affordable housing is greater than the purchase cost of the land. A snapshot of the model is provided in the figure overleaf. Affordable housing feasibility study 5

FIGURE 2: SNAPSHOT OF RLV MODEL Section 0: Base data and assumptions 1 Local Government Area Parramatta_City_of <- Choose LGA from drop down menu 4 Viability threshold 1.25 <- Viable if RLV/current land value is greater than 1.25 5 Valuation of AH includes developer's margin? Yes <- Choose Yes or No (DEFAULT = Yes) 6 Local infrastructure contributions (above Section 94) (LIC) rate/sqm $ - $ per sqm GFA 7 State Infrastructure Charge (SIC) rate/sqm $ - $ per sqm GFA 8 Inclusionary zoning rate (AH as % of res. floor space) 10.0 % 1. Site value with current use 9 Precinct/site Site_6 <- Choose precinct or site from drop down menu 10 Site name Parramatta CBD - 118 Church st 11 Site area 1,727 sqm 12 Current value $ 11,337,426 <- Value from RPData or impute from sale of similar sites 13 Current value/sqm $ 6,565 2. Site value with current zoning ('highest and best use') 14 Precinct/site Site_6_current_zoning <- Choose precinct or site from drop down menu 15 FSR 6.00 :1 16 Gross realisation value (GRV) $ 8,800 per sqm GFA 17 Gross realisation value for residential (GRV) $ 10,000 per sqm GFA 18 Site preparation (demolition/remediation/services) $ 250 per sqm GFA 19 Construction cost $3,300 per sqm GFA 20 Professional fees/contingency (%) 20% per sqm GFA 21 Professional fees/contingency $ 710 per sqm GFA 22 Marketing/financing/holding costs (%) 10% per sqm GFA 23 Marketing/financing/holding costs, etc. $ 880 per sqm GFA 24 Developers margin for profit and risk (% of GRV) 20% per sqm GFA 25 Developers margin for profit and risk per sqm GFA $ 1,760 per sqm GFA 26 Section 94 contributions per sqm GFA $ 128 per sqm GFA 26 Residual Land Value per sqm GFA $ 1,772 (Item 16-18 - 19-21 - 23-25 - 26) 27 RLV per dwelling (based on market dwelling size) $ 142,000 Based on market dwelling size: item 2 28 Permissible floor area based on current FSR 10,362 sqm 29 Residential floor area based on current FSR 6,444 sqm 30 RLV based on current 'highest and best' use $ 18,363,536 31 Viability test (RLV/current land value) w/o AH 1.62 32 Local infrastructure contributions (above Section 94) $ - 33 State Infrastructure Charge (SIC) $ - 34 Affordable housing levy $ 6,444,000 35 RLV based on current 'highest and best' use and AH charges $ 11,919,536 35 Viability test (RLV/current land value) with AH 1.05 3. Site value with upzoning 36 Precinct/site Site_6_new_zoning <- Choose precinct or site from drop down menu 37 FSR 6.61 :1 38 Gross realisation value (GRV) $ 8,900 per sqm GFA 39 Gross realisation value for residential (GRV) $ 10,000 per sqm GFA 40 Site preparation (demolition/remediation/services) $ 250 per sqm GFA 41 Construction cost $3,400 42 Professional fees/contingency (%) 20% per sqm GFA 43 Professional fees/contingency $ 730 44 Marketing/financing/holding costs (%) 10% per sqm GFA 45 Marketing/financing/holding costs, etc. $ 890 46 Developers margin for profit and risk (% of GRV) 20% per sqm GFA 47 Developers margin for profit and risk per sqm GFA $ 1,780 per sqm GFA 48 Section 94 contributions per sqm GFA $ 128 per sqm GFA 49 Residual Land Value per sqm GFA $ 1,722 (Item 38-40 - 41-43 - 45-47 - 48) 50 RLV per dwelling (based on market dwelling size) $ 138,000 Based on market dwelling size: item 36 51 Floor area based on upzoning 11,412 sqm 52 Residential floor area based upzoning 7,494 53 RLV based on upzoning $ 19,653,746 54 Viability test (RLV/current land value) w/o AH 1.73 55 Local infrastructure contributions (above Section 94) $ - 56 State Infrastructure Charge (SIC) $ - 57 Affordable housing levy $ 7,494,000 58 RLV based on upzoning and AH charges $ 12,159,746 59 Viability test (RLV/current land value) with AH 1.07 Land values 60 Current value $ 11,337,426 61 RLV based on current 'highest and best' use $ 18,363,536 62 RLV based on current 'highest and best' use and AH charges $ 11,919,536 63 RLV based on upzoning $ 19,653,746 64 RLV based on upzoning and AH charges $ 12,159,746 65 RLV based on upzoning, VPA and AH charges $ 11,901,704 Source: SGS Economics and Planning, 2017 Affordable housing feasibility study 6

Key parameters The following table summarises the universal assumptions utilised by this model. TABLE 1: UNIVERSAL ASSUMPTIONS AND INPUTS USED BY RLV MODEL Key assumptions Average GFA per dwelling Site preparation costs Professional fees Construction and design contingencies Marketing/financing/holding costs Developer s profit margin Parameters 80 sqm $250 per sqm of GFA 10% of the construction costs 10% of the construction costs 10% of the GRV 20% of the GRV Source: SGS benchmarks. The construction costs and GRV rates vary between different development scenarios, depending on the site locations, proposed uses and built forms. As such, these assumptions are detailed in the following chapter, which discusses the development scenarios being tested on the six sites. Affordable housing feasibility study 7

Note that as site 4 does not have a current FSR control, an average FSR of 2.5:1 is assumed in reflection of the height limit and has been applied to the R4 zoned land to derive a residential yield under the highest and best use scenario. For sites partly or fully zoned B4, the following assumptions have been made with regard to the retail/commercial floorspace on podium level/s: For land zoned B4 on site 1, a retail FSR of 0.5:1 is assumed to provide a ground level of retail floorspace underneath the residential development. For site 6, it is assumed that the commercial floorspace under the highest and best use scenario would be equivalent to the amount proposed under the Planning Proposal (i.e. 3918 sqm of GFA). Further, we consider that the current uses on sites 2 and 3 reflect the highest and best uses under the current controls. As such, no baseline scenario has been developed for these two sites. The following table shows the development yields under both existing and proposed controls for each of the six sites. TABLE 3: DEVELOPMENT YIELDS UNDER EXISTING AND PROPOSED CONTROLS Development yield (sqm of GFA) Site location Granville Rydalmere Melrose Park Epping Carlingford Parramatta CBD Development scenarios Retail Commercial Residential Site_1_current_zoning 1,505 5,670 16,555 Site_1_new_zoning 1,977 27,418 Site_2_new_zoning 18,807 216,281 Site_3_new_zoning 10,000 17,500 392,000 Site_4_current_zoning 31,440 Site_4_new_zoning 2,929 1,384 70,035 Site_5_current_zoning 11,050 Site_5_new_zoning 22,212 Site_6_current_zoning 3,918 6,444 Site_6_new_zoning 3,918 13,352 Source: Developed by SGS based on information provided by Council. 3.2 Revenue and cost assumptions Gross Residual Value per sqm Once the development scenarios have been defined, a rapid market appraisal is completed by SGS to identify an end sales value (or GRV) for each of the proposed uses within these scenarios. In the interests of meeting the short project timeline, this appraisal has been completed mainly through desktop research, which has been focused on compiling the recent sale or rental prices of comparable products in close proximity to the chosen sites. In most cases, there is not sufficient information provided by Council to fully appreciate the quality and unique offerings of the proposed development. As such, professional judgement has been applied in deciding what the recent sales or on-the-market products are mostly comparable to the uses included in the development scenarios, particularly in the case of non-residential uses. The GVR rates shown in the table below reflect what we consider to be most likely end sale price of each proposed use under the current market conditions. Affordable housing feasibility study 9

TABLE 4: ESTIMATED GRV PER SQM OF GFA Site location Granville Rydalmere Melrose Park Epping Carlingford Parramatta CBD Development scenarios GRV per sqm Retail Commercial Residential Average GRV per sqm Site_1_current_zoning $5,400 $5,300 $8,200 $7,300 Site_1_new_zoning $5,400 $8,200 $8,000 Site_2_new_zoning $3,000 $7,900 $7,500 Site_3_new_zoning $6,300 $3,000 $9,400 $9,000 Site_4_current_zoning $11,500 $11,500 Site_4_new_zoning $4,700 $6,700 $11,500 $11,100 Site_5_current_zoning $8,100 $8,100 Site_5_new_zoning $8,100 $8,100 Site_6_current_zoning $6,700 $10,000 $8,800 Source: SGS estimates, 2017 Site_6_new_zoning $6,700 $10,000 $8,900 Note the last column in the table above shows the average GRV per sqm for each scenario used in the RLV model. This is calculated as a weighted average of the GRVs of the proposed uses, while taking in account the mix of development yield shown in Table 3. Construction costs per sqm The per-sqm construction cost rates used in this analysis are sourced from the Rawlinson Construction Handbook 2017. The selected rates for each scenario are displayed in the table below. TABLE 5: CONSTRUCTION COSTS PER SQM Site location Granville Rydalmere Melrose Park Epping Carlingford Parramatta CBD Development scenarios Construction cost per sqm Retail Commercial Residential Average cost per sqm Site_1_current_zoning $2,038 $3,336 $3,879 $3,600 Site_1_new_zoning $2,038 $4,009 $3,900 Site_2_new_zoning $2,816 $3,879 $3,800 Site_3_new_zoning $2,038 $2,816 $3,879 $3,800 Site_4_current_zoning $3,879 $3,900 Site_4_new_zoning $2,038 $2,816 $3,879 $3,800 Site_5_current_zoning $3,879 $3,900 Site_5_new_zoning $4,009 $4,000 Site_6_current_zoning $2,568 $3,810 $3,300 Site_6_new_zoning $2,568 $3,810 $3,400 Source: SGS calculations using per sqm rates from Rawlinson Construction Handbook 2017 Note the rates for commercial and residential development above take into account the cost of constructing required underground parking per sqm of GFA. For example, according to the Parramatta LEP, 1.2 car spaces (including visitor parking) are required for every dwelling developed in the Parramatta CBD. Based on 30 sqm per car space and 80 sqm of GFA per apartment, every sqm of residential GFA needs to provide 0.45 sqm (30/80*1.2) of underground car parking, which costs approximately $1768 per sqm or $795 per sqm of residential GFA. Adding this cost to the per-sqm construction cost of around $3015 for high- Affordable housing feasibility study 10

rise apartment gives a total cost of $3810 per sqm for residential development, as shown in the last row of the table above. For sites outside the CBD, a different parking provision rate of 1.5 spaces per dwelling is used to calculate the total construction cost. In addition to the variation in parking provision rate, the variation in the per sqm cost also reflects the built form assumed for each scenario. For example, a higher rate is used in the Site_1_new_zoning scenario, because the residential development is likely to exceed 9 storeys, which has higher structure cost, compared to the development (below 9 storeys) under the Site_1_current_zoning scenario. As with the GRV rates, a weighted average construction cost (in the last column) is calculated for each scenario and is inputted into the RLV model. Infrastructure contributions S.94/S.94A contributions S94 or S.94A contributions have been calculated for each scenario, using the following rates prescribed in Council s development contributions plans: Epping o $184.99 per sqm of retail GFA o $42.71 per sqm of commercial GFA o $250 per sqm of residential GFA Parramatta CBD 3% of the development costs Rest of the Parramatta LGA 1% of the development costs. VPA On top of these contributions, we note that a VPA offer valued at $89.8 million has been received by Council for the Melrose Park site. Of this amount, $31.7 million is associated with the on-site provision of new parks and open space, as well as provision of community and child care facility. As we expect that the $31.7 million is largely made up of the land value for dedicated open space and construction costs of the community and child care facility (which have already been taken into account in the RLV analysis), we have deducted that amount from the $89.8 VPA offer to avoid potential double-dipping. This returns a remaining VPA offer of $58.1 million for provision of off-site infrastructure. Similarly, the VPA offer received for the Epping site has not been included in this analysis. This is because the land dedication for public open space as part of this offer does not add costs to the development equation, when the land acquisition cost is already included in the RLV analysis. However, the other parts of the VPA offer, including embellishment and maintenance of open space, as well as contributions towards a community facility, will result in additional costs to the developer, but has not been included in this analysis due to lack of costings for these items. Value capture policy In addition to local infrastructure contributions, a value capture charge is applied to the Parramatta CBD site. This charge is $150 per sqm of additional residential floorspace up to an incentive FSR level. This has been derived from 20% of the land value uplift associated with the additional FSR above the current control. Special infrastructure contribution According to a media release by Transport for NSW back in 2015, a special infrastructure contribution (SIC) will be implemented along the Parramatta Light Rail Corridor, with the levy expected to be set at around $200 per sqm of new residential floorspace subject to consultation. Note this feasibility analysis has not included the SIC levy, given the uncertainty around when this levy may come into place and its actual amount. Further, the sales price assumptions used in this study is based on a current market appraisal of the proposed uses, and therefore may not have taken into account any inflationary impact that the Light Rail project will have Affordable housing feasibility study 11

on sales price. As such, including SIC in this testing would exaggerate its impact on development feasibility. ARH contributions As per requirements in the brief, the ARH contributions have been calculated for each scenario under two options: Option 1: 10% (or 30% on government-owned land) of the dwellings handed over to Council at no charge - In this case, the contributions are calculated as 10% (or 30% on government-owned land) of the residential GFA valued at the market GRV rates mentioned above. Option 2: 10% of the residential GFA handed over to Council with an at-cost charge the contributions are calculated as the difference between the value under option 1 and costs of constructing the 10% of the dwellings as ARH. Purchase cost of land The purchase price paid for each site has been extracted from historical sales records provided by RPdata, except from site 2 which is owned by the government. These prices are provided in the table below. TABLE 6: PURCHASE COST OF LAND Site no. Site name Purchase price 1 Granville $6,500,000 2 Rydalmere $53,400,000 (as per VG estimate) 3 Melrose Park $144,500,000 4 Epping $64,659,833 5 Carlingford $9,628,594 6 Parramatta CBD $11,337,426 Source: SGS calculations based on RPData sales record Note the price of site 2 has been based on the Valuer General estimate, whereas the price of site 6 has been inflated from the last sale price of $5.9 million in 2007 to 2015 (when most of the other sites were bought), using the capital growth rate for houses in Parramatta suburb over the intervening period. As discussed earlier, the development is deemed feasible, when the RLV after deducting all the development costs including ARH contributions is greater than the purchase price of the land (remembering that in some of these cases the purchase price of the land may have included a premium or speculative component in anticipation of a future upzoning). Affordable housing feasibility study 12

4. FEASIBILITY RESULTS Using the inputs and assumptions discussed earlier, a RLV is calculated for each scenario with and without the ARH charge. This chapter presents the key results and findings from the feasibility assessment. 4.1 Key results Feasibility of handing over ARH at no charge The table below provides the RLVs for each site under both the current and proposed controls. These results are based on the assumption that the ARH dwellings would be provided to Council at no charge. TABLE 7: KEY RESULTS OF THE FEASIBILITY ASSESSMENT 10% ARH CONTRIBUTIONS AT NO CHARGE Granville Rydalmere Melrose Park Epping Carlingford Parramatta CBD 1 Purchase cost of land $6.5M $53.4M 1 $144.5M $51.3M $9.6M $11.3M 2 3 4 RLV based on current 'highest and best' use RLV based on current 'highest and best' use and ARH contribution Viability test under current controls with ARH contribution (i.e. item 3/ item 1) $10.5M $53.4M 1 $21M 2 $88.7M $7.1M $18.4M -$3.0M N/A N/A $52.5M -$1.9M $11.9M -0.47 N/A N/A 1.02-0.19 1.05 5 RLV based on upzoning $16.9M $80.3M $583.7M $197.8M $11.6M $32.5M 6 7 8 RLV based on upzoning and ARH contribution RLV based on upzoning, VPA/value capture and ARH contribution Viability test under new controls with ARH contr (i.e. item 7 or 6/ item 1) -$5.6M -$432.3M $215.2M $117.2M -$6.4M $19.2M N/A N/A $157.1M N/A N/A $16.6M -0.86-8.10 1.09 2.28-0.67 1.46 Source: SGS calculations, 2017 1. Since the Rydalmere site is owned by the government, no sale record exists in RPData. As such, the VG estimate is used as a proxy for the price that would be paid for the site, or the current underlying land value (i.e. RLV under the current highest and best use) 2. We believe the current use on the Melrose Park site reflects the highest and best use under current zoning and controls. According to the VG estimate, the site was worth $21 million in 2015 before it was revalued to $100 M by VG following the sale of site for $144 M in Aug 2015. Therefore, $21 million is used as a proxy for the current underlying value under the existing zoning and controls. Two viability tests have been performed. The first one calculates the ratio of the RLV after the ARH contributions under the current highest and best use scenario, compared to the purchase price paid for the land. This test found that the 10% ARH target is only feasible for the Epping and Parramatta CBD sites under the current controls. In the cases of Granville and Carlingford, the 10% ARH contribution is not viable and would represent a cost impost to the development under the current controls. The second test calculates the ratio of the RLV after the ARH contributions (in addition to any VPA and value capture charges) under the new planning controls sought by the developer, compared to the purchase price for the land. This test found that the 10% ARH contribution becomes much more feasible for the Epping site, with the additional residential yield sought by the Planning Proposal. Also, the 10% ARH target is still feasible for the Parramatta site under the higher FSR scenario, even with a value capture charge levied at $150 per sqm of the bonus floorspace. Affordable housing feasibility study 13

The Melrose Park development under the Planning Proposal can also meet a 10% ARH target, after a VPA contribution valued by the developer at $89.8M. On the other hand, the Planning Proposal for the Rydalmere site would not be able to meet the 30% ARH target proposed by CoP for government-owned land. This is because the development only generates a RLV of around $80M without the ARH contributions, which is greater than the current valuation of the land but not enough to give away 30% of the new dwellings for free. In practice a 30% ARH contribution on the government owned land would require the government to sell the land at less than its currently estimated value. With the upzoning, the 10% ARH contribution is still not viable on the Granville and Carlingford sites. This can be expected as the ARH levy was not anticipated by the developer who bought these sites without factoring in the costs of providing ARH at no charge. Interestingly, the modelling also suggests that the proponents for the Carlingford and Epping site paid more than what these sites would be worth under the current highest and best use scenario (i.e. item 2 in the table above), in anticipation of a future upzoning. Generally, the price paid for the development site should be less than or comparable to the current underlying land value, in absence of a speculation on additional development potential. Feasibility of handing over ARH at cost As an alternative option, the feasibility impact of providing ARH at cost is also examined. Under this option, the developer will be producing the required ARH to cover costs, therefore these ARH dwellings would not have any impact on the RLV. In other words, this option effectively compresses the saleable residential yield by the % required for ARH across all development scenarios. TABLE 8: KEY RESULTS OF THE FEASIBILITY ASSESSMENT 10% ARH CONTRIBUTIONS AT COSTS Granville Rydalmere Melrose Park Epping Carlingford Parramatta CBD 1 Purchase cost of land $6.5M $53.4M $144.5M $51.3M $9.6M $11.3M 2 3 4 RLV based on current 'highest and best' use RLV based on current 'highest and best' use and AH charges Viability test under current controls with AH (i.e. item 3/ item 1) $10.5M $53.4M $21M $88.7M $7.1M $18.4M $9.8M N/A N/A $79.8M $6.4M $17.6M 1.51 N/A N/A 1.55 0.66 1.56 5 RLV based on upzoning $16.9M $80.3M $583.7M $197.8M $11.6M $19.7M 6 7 8 RLV based on upzoning and AH charges RLV based on upzoning, VPA/value capture and AH charges Viability test under new controls with AH (i.e. item 7 or 6/ item 1) Source: SGS calculations, 2017 $15.3M $46.2M $529.1M $184.4M $10.4M $28.9M N/A N/A $471.0M N/A N/A $27.1M 2.35 0.87 3.26 3.59 1.08 2.39 Compared to providing ARH for free, this option has a much smaller impact on development feasibility. As can be seen from the results above, all developments under the new controls would be viable with provision of 10% ARH at cost. With provision of 30% ARH at cost, the Rydalmere site can be sold for up to $46 million, which is $6 million less than the VG valuation. Under the current controls, the 10% ARH contribution would still be viable, except for the Carlingford site, where the RLV under the current residential FSR is less than the purchase price paid for the site. Affordable housing feasibility study 14

4.2 Additional ARH scenarios Following the feedback from Council staff on the draft report, we have been asked to test a number of additional ARH scenarios as follows: Private sites Government site 5% of total project dwelling yield at no cost 8% of total project dwelling yield at no cost 15% of total project dwelling yield at no cost 10% of the value uplift from the rezoning/upzoning as ARH contributions 15% of the value uplift from the rezoning/upzoning as ARH contributions 20% of the value uplift from the rezoning/upzoning as ARH contributions A hybrid ARH option is also added, which comprises a 5% levy of total dwelling yield under the current zoning/controls and 10% of value uplift from the rezoning. The following section discusses the feasibility results for the additional scenarios described above. 8% of total project dwelling yield at no cost The table below shows the feasibility results on five private sites, if 8% of the total dwelling yield is handed to Council as ARH at no cost. The Granville and Carlingford sites remain unfeasible, after reducing the ARH requirement from 10% to 8% of the total dwelling yield. TABLE 9: KEY RESULTS OF THE FEASIBILITY ASSESSMENT 8% ARH CONTRIBUTION AT COST Granville Melrose Park Epping Carlingford Parramatta CBD 1 Purchase cost of land $6.5M $144.5M $51.3M $9.6M $11.3M 2 3 4 RLV based on current 'highest and best' use RLV based on current 'highest and best' use and ARH contribution Viability test under current controls with ARH contribution (i.e. item 3/ item 1) $10.5M $21M 2 $88.7M $7.1M $18.4M -$0.3M N/A $59.7M -$0.1M $13.2M -0.05 N/A 1.16-0.01 1.17 5 RLV based on upzoning $16.9M $583.7M $197.8M $11.6M $32.5M 6 7 8 RLV based on upzoning and ARH contribution RLV based on upzoning, VPA/value capture and ARH contribution Viability test under new controls with ARH contr (i.e. item 7 or 6/ item 1) Source: SGS calculations, 2017 -$1.1M $288.9M $133.3M -$2.8M $21.8M N/A $230.8M N/A N/A $19.2M -0.17 1.60 2.60-0.29 1.70 5% of total project dwelling yield at no cost The table below shows the feasibility results on five private sites, if 5% of the total dwelling yield is handed to Council as ARH at no cost. After reducing the ARH requirement to 5% of the total dwelling yield, the Carlingford site remains unfeasible whereas the Granville site becomes marginally feasible with a feasibility ratio of 0.87. Affordable housing feasibility study 15

TABLE 10: KEY RESULTS OF THE FEASIBILITY ASSESSMENT 5% ARH CONTRIBUTIONS AT COSTS Granville Melrose Park Epping Carlingford Parramatta CBD 1 Purchase cost of land $6.5M $144.5M $51.3M $9.6M $11.3M 2 3 4 RLV based on current 'highest and best' use RLV based on current 'highest and best' use and ARH contribution Viability test under current controls with ARH contribution (i.e. item 3/ item 1) $10.5M $21M 2 $88.7M $7.1M $18.4M $3.7M N/A $70.6M $2.6M $15.1M 0.58 N/A 1.37 0.27 1.34 5 RLV based on upzoning $16.9M $583.7M $197.8M $11.6M $32.5M 6 7 8 RLV based on upzoning and ARH contribution RLV based on upzoning, VPA/value capture and ARH contribution Viability test under new controls with ARH contr (i.e. item 7 or 6/ item 1) Source: SGS calculations, 2017 $5.6M $399.5M $157.5M $2.6M $25.8M N/A $341.4M N/A N/A $23.2M 0.87 2.36 3.07 0.27 2.05 15% of total project dwelling yield at no cost on government site The table below displays the feasibility results for the government site in Rydalmere, if 15% of the project dwelling yield is handed back to Council as ARH at no cost. As can be seen below, the development remains unfeasible, because the ARH contribution levied at 15% of the total project dwelling yield is greater than the RLV of the proposed development. TABLE 11: RESULTS OF THE FEASIBILITY ASSESSMENT - 15% ARH CONTRIBUTION AT NO COST ON GOVERNMENT SITE Rydalmere - Government site 1 Purchase cost of land $53.4M 2 RLV based on current 'highest and best' use $53.4M 3 4 RLV based on current 'highest and best' use and ARH contribution Viability test under current controls with ARH contribution (i.e. item 3/ item 1) N/A N/A 5 RLV based on upzoning $80.3M 6 RLV based on upzoning and ARH contribution -$176.0M 7 8 RLV based on upzoning, VPA/value capture and ARH contribution Viability test under new controls with ARH contribution (i.e. item 7 or 6/ item 1) Source: SGS calculations, 2017 N/A -3.30 10% of the value uplift as ARH contributions The table below displays the feasibility results for five private sites, if 10% of the value uplift from rezoning or upzoning is levied as an ARH contribution. In this case, the ARH contribution is a proportion of the uplift in land value the developer would be enjoying if the additional development right is granted. As such, the levy would not impact on the development feasibility, as shown below. An affordable housing yield has also been calculated by dividing the monetary contribution by the average sale price of new dwellings (it should be noted that by working with a Community Housing Provider willing to purchase dwellings at cost, the feasibility equation might be changed for a developer and additional dwellings may be able to be provided compared to Affordable housing feasibility study 16

those assumed at market sale prices). The Melrose Park site can generate 51 ARH dwellings, whereas the relatively smaller site can only contribute one ARH dwelling. TABLE 12: RESULTS OF THE FEASIBILITY ASSESSMENT 10% OF THE VALUE UPLIFT AS ARH Granville Melrose Park Epping Carlingford Parramatta CBD 1 Purchase cost of land $6.5M $144.5M $51.3M $9.6M $11.3M 2 RLV based on current 'highest and best' use $10.5M $144.5M $88.7M $7.1M $18.4M 3 RLV based on upzoning and VPA/value capture charge $16.9M $525.6M $197.8M $11.6M $29.9M 4 Value uplift 1 (i.e. item 3 - item 2) $6.3M $381.1M $109.1M $1.9M $11.5M 5 ARH contributions ($) (i.e. 10%*item 4) $0.6M $38.1M $10.9M $0.2M $1.2M 6 ARH contributions (dwellings) 1 51 12 0.3 1 7 8 RLV based on upzoning and AH contributions Viability test (RLV/purchase cost of land) with AH (i.e. item 7 / item 1) $16.2M $487.5M $186.9M $11.4M $28.7M 2.50 3.37 3.64 1.18 2.54 Source: SGS calculations, 2017 1. Value uplift has been calculated as the difference between the RLVs under the current and proposed controls after any VPA or value capture charges. Except for Carlingford, it was calculated as the difference between the purchase price of land and RLV under the proposed controls. 15% of the value uplift as ARH contributions The table below displays the feasibility results for five private sites, if 15% of the value uplift from rezoning or upzoning is levied as ARH contribution. For the reason explained above, the levy would not impact on the development feasibility as shown below. Under this option, the affordable housing yield on the Melrose Park site increases to 76 dwellings. The Epping site can also contribute around 18 ARH dwellings. TABLE 13: RESULTS OF THE FEASIBILITY ASSESSMENT 15% OF THE VALUE UPLIFT AS ARH Granville Melrose Park Epping Carlingford Parramatta CBD 1 Purchase cost of land $6.5M $144.5M $51.3M $9.6M $11.3M 2 RLV based on current 'highest and best' use $10.5M $144.5M $88.7M $7.1M $18.4M 3 RLV based on upzoning and VPA/value capture charge $16.9M $525.6M $197.8M $11.6M $29.9M 4 Value uplift 1 (i.e. item 3 - item 2) $6.3M $381.1M $109.1M $1.9M $11.5M 5 ARH contributions ($) (i.e. 10%*item 4) $1.0M $57.2M $16.4M $0.3M $1.7M 6 ARH contributions (dwellings) 1 76 18 0.4 2 7 8 RLV based on upzoning and AH contributions Viability test (RLV/purchase cost of land) with AH (i.e. item 7 / item 1) $15.9M $468.4M $181.4M $11.3M $28.2M 2.45 3.24 3.53 1.17 2.48 Source: SGS calculations, 2017 2. Value uplift has been calculated as the difference between the RLVs under the current and proposed controls after any VPA or value capture charges. Except for Carlingford, it was calculated as the difference between the purchase price of land and RLV under the proposed controls. 20% of the value uplift as ARH contributions The table below displays the feasibility results for five private sites, if 20% of the value uplift from rezoning or upzoning is levied as ARH contribution. For the reason explained above, the levy would not impact on the development feasibility as shown below. Affordable housing feasibility study 17

Under this option, just over 100 ARH dwellings can be generated by the Melrose Park development, whereas the Epping site can contribute up to 24 dwellings. TABLE 14: RESULTS OF THE FEASIBILITY ASSESSMENT 20% OF THE VALUE UPLIFT AS ARH Granville Melrose Park Epping Carlingford Parramatta CBD 1 Purchase cost of land $6.5M $144.5M $51.3M $9.6M $11.3M 2 RLV based on current 'highest and best' use $10.5M $144.5M $88.7M $7.1M $18.4M 3 RLV based on upzoning and VPA/value capture charge $16.9M $525.6M $197.8M $11.6M $29.9M 4 Value uplift 1 (i.e. item 3 - item 2) $6.3M $381.1M $109.1M $1.9M $11.5M 5 ARH contributions ($) (i.e. 10%*item 4) $1.3M $76.2M $21.8M $0.4M $2.3M 6 ARH contributions (dwellings) 2 101 24 1 3 7 8 RLV based on upzoning and AH contributions Viability test (RLV/purchase cost of land) with AH (i.e. item 7 / item 1) $15.6M $449.4M $175.9M $11.2M $27.6M 2.40 3.11 3.43 1.16 2.43 Source: SGS calculations, 2017 3. Value uplift has been calculated as the difference between the RLVs under the current and proposed controls after any VPA or value capture charges. Except for Carlingford, it was calculated as the difference between the purchase price of land and RLV under the proposed controls. 5% of total dwellings under current controls + 10% of value uplift As discussed above, this is a hybrid option which comprises a 5% levy of total dwelling yield under the current zoning/controls and 10% of value uplift from the rezoning. As shown below, this option would be feasible on most of the five private sites tested, with only the Carlingford site found marginally unviable, while generating significantly more ARH dwellings compared to a 20% value uplift levy. TABLE 15: RESULTS OF THE FEASIBILITY ASSESSMENT 5% OF TOTAL DWELLINGS UNDER CURRENT CONTROLS + 10% OF VALUE UPLIFT AS ARH Granville Melrose Park Epping Carlingford Parramatta CBD 1 Purchase cost of land $6.5M $144.5M $51.3M $9.6M $11.3M 2 RLV based on current 'highest and best' use $10.5M $144.5M $88.7M $7.1M $18.4M 3 RLV based on upzoning $16.9M $525.6M $197.8M $11.6M $29.9M 4 ARH contributions ($) - 5% of the dwellings under current zoning $6.8M $184.2M $18.1M $4.5M $3.2M - 10% of the value uplift $0.6M $38.1M $10.9M $0.2M $1.2M 5 ARH contributions (dwellings) 11 296 32 7 5 6 7 RLV based on upzoning and AH contributions Viability test (RLV/purchase cost of land) with AH (i.e. item 6 / item 1) Source: SGS calculations, 2017 $9.4M $303.2M $168.8M $6.9M $25.5M 1.45 2.10 3.29 0.72 2.25 Affordable housing feasibility study 18

5. CONCLUSION The feasibility testing completed for this research report has found a varying impact of the ARH contributions on different sites within the Parramatta LGA. The analysis represents a robust test of the different possible ARH contributions but the sites and developments considered are only a selection based on current proposals, and may not necessarily be representative of the range of development circumstances across the LGA. The results for the Carlingford site in particular, where regular non-feasible results were returned, seem at odds with what might be expected, indicating that the combination of the underlying existing site value, construction costs and sale values used in the analysis may not reflect what is achievable at other sites in this locality. In addition, the Parramatta Light Rail (and any future Special Infrastructure Contribution) is likely to have an inflationary impact on achievable sale values in the corridor it will serve, but best estimates for current achievable values are utilised in the modelling (and for similar reasons no SIC is assumed). In addition, it should be noted that the feasibility analysis has been completed without the benefit of a Quantity Surveyor (QS) input on costs, and the revenue assumptions utilised by this analysis have been developed based on a rapid market appraisal conducted largely through desktop research. It is recommended that Council engage the services of QS and a valuer to cross check the cost and revenue assumptions, should Council wish to utilise this work for more detailed negotiations on ARH contributions with the proponents for the sites and Planning Proposals considered in this report. Key findings and policy implications from the site-specific feasibility analysis are summarised in the table below: Key findings from the feasibility testing Development proposals seeking a rezoning or significant uplift in residential yield from the current controls are more likely to meet the required ARH target at no charge while keeping the overall development feasible, compared to the redevelopment under the current controls. Policy implications The implication for the Affordable Housing policy is that a 10% ARH contribution at no charge is mostly likely to be acceptable where it applies to land being up-zoned (i.e. the planning controls are being amended or modified to allow additional development potential). The extent that any development will remain feasible after a contribution of affordable housing will often depend on the additional density or value uplift that is being generated (and whether the site was purchased at a value that reflected its pre-rezoning potential). The provision of 10% ARH at no cost tends to be not feasible for developments, where the residential GRV is lower than $9000 per sqm (i.e. Granville and Carlingford) or the site was bought by the developer for a price much higher than what it would be worth with the current zoning and controls. These sites remain unfeasible, after reducing the % of the dwelling requirement to 5%. A premium price is often paid for the development site when there is a speculation on an upzoning or rezoning, which would result in a significant uplift in the land value. This highlights that there may be some sub-markets where it is more difficult to make a free affordable housing contribution work from a feasibility perspective, while still satisfying environmental and planning standards. It is possible that a more modest percentage target for affordable housing should apply as the inclusionary standard. Affordable housing contributions beyond this could be based on a share of the uplift in land value, as is being contemplated for the value capture charge in the CBD. This approach was tested in the hybrid example shown in Table 15 and found to be feasible (at the percentages tested) in all cases except for the Carlingford site (see discussion below). The implication for the policy is that its general application should not be retrospective, particularly if a higher percentage ARH contribution is expected. In other Affordable housing feasibility study 19

Under these cases, an ARH levy requiring a % of the dwelling yield is more likely to be unviable, because such a charge was not anticipated by the developer in the price for the land. The provision of 30% ARH at no cost on government land in Rydalmere is found not feasible, because the uplift in land value from the rezoning is not high enough to hand over 30% of dwellings as affordable housing for free. However, the Rydalmere development can provide 30% ARH at cost if the site is sold to the developer at a discount of around 10%. The development remains unfeasible, after reducing the % of the dwelling requirement to 15%. The provision of 10% ARH at cost is found feasible for almost all of the development scenarios tested. A levy capturing up to 20% of the value uplift from rezoning has little impact on development feasibility, but this option would only generate a modest amount of ARH dwellings particularly on small sites. A hybrid option of 5% ARH requirement of total dwellings under the current controls plus 10% of value uplift from the additional development right is found feasible across most of the private sites tested. words, it is best to apply to future Planning Proposals or sites which are sold after the policy has been adopted. Goodwill negotiations on sites not in this category (i.e. sold more recently in anticipation of a change to development controls) could still proceed with the policy as a guide to desirable outcomes. The implication is that the government may have to subsidise the development by selling the land at a discounted price, compared to what it might otherwise receive given the development controls, in order to make the provision of a high percentage (such as 30%) of ARH on government land. This may still be desirable, but the implication of the effective subsidy needs to be recognised. An at cost ARH contribution consistent with the proposed policy could probably apply to all Planning Proposals going forward, without having regard to when the subject site was purchased, provided the proposed additional density (FSR) for residential development is greater than 10%. By definition, a value capture charge would not impact on development feasibility. Typically, 50% of value uplift is suggested as a realistic (though not uncontested) upper limit for such a charge to fund public benefits. If applied to all Planning Proposals at this rate the ARH contribution would be considerable (it could be provided in cash for equivalent ARH dwellings). However, there are other public benefit claims for value capture levies (including transport and other infrastructure, for example as proposed in the CBD) so it is probably unrealistic for a 50% levy to apply solely to provide ARH. This finding reinforces the policy implication discussed above regarding a hybrid approach. Specifically, it suggests that a modest percentage of dwelling yield would be anticipated as an inclusionary ARH contribution under the current controls, while proponents seeking additional development rights would pay an additional ARH levy based on a share of the uplift in land value gained from upzoning. This option is likely to be more effective in terms of generating a base quantum of ARH dwellings through the inclusionary component, compared to relying on a value capture levy applying to Planning Proposals and rezonings only. The above conclusions, based on the feasibility analysis in this report, do not constitute recommendations for Council s Affordable Housing Policy. The findings and implications need to be considered - and the final policy settings decided (i.e. ARH based on an inclusionary standard, a share of value capture or a hybrid approach) - by reference to Council s Affordable Housing Discussion Paper and in particular the ARH target of 9,500 dwellings. Given the anticipated amount of development going forward and the additional land value created it would be possible to create a closer link between the identified need and the Affordable housing feasibility study 20