Exploring Conventional Economic Viability as a Potential Barrier to Scalable Urban Agriculture: Examples from Two Divergent Development Contexts

Abstract

Urban Agriculture (UA) is the widespread practice of food production within available city space using non-commercial, commercial and hybrid production technologies. The economic viability of UA remains a concern among UA practitioners. To investigate UA’s viability; land, labour and distribution cost are analyzed, and margin and benefit–cost ratio (BCR) under vacant lot, rooftop/backyard and discretionary labour UA are calculated. We present a straightforward approach to gauge the economic viability of UA taking examples from 40 distinct locations of two divergent development contexts of Adelaide, South Australia and Kathmandu Valley, Nepal. UA seems potentially viable by selecting high-value crops in Adelaide but showed little chance of viability under low-value crop scenarios in both contexts. The high cost of land is shown to be the primary driver of cost for UA. Labour cost appears to be a critical difference between the two cities, being an important constraint for the economic viability in Adelaide, where the wage rate is high. To improve economic viability, the respective governments and planners should consider better ways to avail subsidised land through policy intervention and volunteer or subsidised labour arrangement mechanisms. Home food gardens accessing available land and labour as a discretionary/spare time activity with zero distribution cost may represent the best way to produce food without exceeding market costs in cities.

1. Introduction

The global human population is increasing, and more than half of the global population lives in urban areas, which is expected to rise to 68% by 2050. The problem of food and nutritional security can be severe in urban centres due to the disruption of the food supply chain, physical and economic barriers and food waste due to labour shortages, including from pandemics like COVID-19 [1]. After the COVID-19 pandemic, public interest in homegrown fruits and vegetables has soared for improved resilience [2]. Megacities are pushing for a more resilient food system that enhances local production, and Urban Agriculture (UA) is promoted as one of the best strategies to cope with the present barriers to food production [1]. UA is a food production system undertaken within urban areas to provide safe and diverse food products and potentially save or earn money using non-commercial, commercial and hybrid production technologies [3].

Diverse in its mission, scale, means, and forms [4], UA can provide some portions of fresh and safe food with shorter and simpler supply chains than the conventional food system, with other claimed social and environmental benefits [5]. A typical characteristic of UA is the combination of small and dispersed production units creating a supply system mainly within the proximity of consumption [6]. The primary UA practices include community gardens, commercial farms, rooftop farming, community-supported agriculture, and indoor farms as the recent advanced practices [7]. UA can involve production for self-consumption and/or sale and is promoted based on integrated social, economic, and environmental benefits [8].

Several drivers have fostered the expansion of UA in the past, including war [9], coping with economic crises, environmental protection [10], and a growing interest in developing a self-sufficiency economy [11]. In addition, due to its potential contribution to reducing food costs, UA practices have increased in cities [12].

If UA is to deliver improved food security, then the economic viability should be of concern, explicitly or implicitly, to different stakeholders. Viability relates to the efficiency of resources used for production at a given time. Costs and benefits are typically used to measure economic viability [13], but viability can be defined differently. Profitability, liquidity, stability, and productivity are common indicators for measuring economic viability [14]. Revenue and cost are essential components for computing profitability; the availability of cash in hand determines the state of liquidity. The amount of capital gain decides stability, while the availability of input resources determines output as the measure of productivity [14]. In a general sense, long-term viability governs the sustainability of any enterprise [15]. Therefore, economic viability and impacts in the urban area can be viewed as the primary criteria underpinning the financial sustainability of UA [16].

UA is increasingly viewed implicitly as being economically viable, as it is expected to contribute to the social, economic, and environmental objectives of sustainable urban development [17]. According to Clinton et al. [18], proper implementation of UA at a global scale can save up to 160 billion USD per year through food production and ecosystem services. A study on the economic viability of UA has focused on multifunctional benefits from UA, considering direct and quantifiable benefits (cash revenue from production) plus indirect and less quantifiable social and health benefits [19]. Despite different claims, there are limited studies on the direct economic benefits of UA considering conventional factors like profitability, liquidity, and productivity [14], and most of the available studies omit typical agricultural production input costs such as land, labour, and distribution [20]. Amidst growing support for UA, surprisingly little has been explored in terms of its economic benefits [21]. The efficacy of urban farming for saving money is not well documented, and the potential savings are mainly calculated based on the gross retail value of produce without valuing the full input costs [12].

UA is prevalent in both high and low-income countries. UA is rapidly expanding in high-income countries like Australia, Canada, the United States, England, France, and New Zealand. Despite the shortened supply chain, UA is potentially expensive due to high initial investment, soaring land costs, and increasing labour and distribution costs [22]. The availability of affordable and usable land in urban areas and its cost is a primary challenge for prospective urban growers to achieve financial viability [23]. Maiti [24] reiterated that land cost, lack of capital resources for investment and limited technical assistance are the most prominent challenges to the viability of UA in cities. The cost of establishing and maintaining a food garden may plausibly be greater than buying food due to capital and the labour invested in year-round production. If UA is to deliver affordable improvements to food security, particularly for low-income households, research is warranted to investigate the cost-effectiveness of this mode of food production and distribution [25].

Most earlier researchers in UA have not investigated the total UA production cost, and in most cases, they did not account for the cost of land, labour, and distribution. The high price of land, limited space for production expansion and high living costs, combined with potentially low availability of input resources and lack of efficiency at small production scales, represent significant challenges in the quest for profitable UA in most cities [17].

The objective of this study is to carry out a ‘reality check’ on the fundamentals of a viable UA business model focusing on three common conventional UA practices (vacant lot, backyard/rooftop and discretionary labour) based on the hypothetical scenario in which UA is scaled up to the point where competitive market rates must be paid to access land and labour and to distribute produce. Using this simplified analysis, we explore possible options for returning an acceptable wage rate for the above three conventional UA types and the significance and impact of distribution costs under two divergent developmental contexts of Adelaide, South Australia, Australia and Kathmandu Valley, Nepal.

2. Materials and Methods

2.1. Study Locations

The greater Adelaide region is a sprawling city in South Australia, Australia. It covers 3260 km² with a population of 1.3 million people at an average density of 400 people per square kilometer, and with a large population concentrated in small suburbs [26]. UA is prevalent in Australia—in the states of South Australia, Tasmania, and Victoria, 59% of households are involved in South Australia [27]. Home food gardens are the most common in Australia, with little exploration of productivity and financial savings [28]. Both private and public level UA is prevalent in Adelaide [29].

Kathmandu Valley includes Kathmandu, Bhaktapur and Lalitpur cities, with a total area of 570 km² [30]. It is one of the most crowded cities in South Asia, with a population of 2.54 million, which is growing by 6.5% per year [31]. Within three decades, the rate of urbanisation in Kathmandu Valley has increased by 412%, with a loss of 31% of productive agricultural land [32]. UA in Kathmandu Valley is sprawling due to its fertile soil and a high degree of public interest in growing food, coupled with the increased demand for food from a rapidly growing urban population [33].

UA practices are based on conventional soil-based farming in both Adelaide and Kathmandu Valley. Farming in the available lot of land with a front or backyard garden is the most common form of UA in Adelaide [29]. In contrast, based on the lead author’s experience, rooftops/backyards, and soil-based conventional farming in available lots are typical in Kathmandu Valley.

2.2. Conceptual Frame

Most previous research evaluating the economic benefits of UA has neglected the full cost of production, including land, labour and distribution [20]. This study analyses the impact of land, labour, and distribution costs on the economic viability of UA based on plausible estimates for cost, crop productivity, crop value, and price scenarios under vacant lot, backyard/rooftop and discretionary labour UA using a conceptual economic framework (Figure 1). A crude range of outcomes for potential economic viability is estimated by taking examples of land, labour, and distribution costs within different distances and scales from urban centres, based on a review of available land parcels for sale.

The economic viability of UA is estimated based on cost and return per square metre expressed as Net Margin (NM) and Benefit Cost Ratio (BCR) under the given data-informed scenarios of land, labour and distribution costs with different production and crop value:

Zero and non-zero land costs;

Zero and non-zero labour costs;

Zero and non-zero distribution costs;

High-value and low value crops; and

High production, base production, and low production.

2.3. Methodology

The method adopts a deliberately parsimonious approach to simulate the income and costs in UA based on typical parameters using an economic model. The model is based on developing overarching viability based on determining the net margin and Benefit Cost Ratio, which we call NM ($/yr) and BCR, respectively. NM is calculated under the three labour/yield scenarios (high labour/high yield, base labour/base yield, and low labour/low yield) for the three UA types (vacant lot, backyard/rooftop, and discretionary labour/hobby) and is given by:

NM = I − C_total,

(1)

where:

I = Gross income ($/yr)

C_total = total costs ($/yr)

Similarly, for conventional economic viability, BCR is calculated, and a business is considered viable if the BCR is greater than one and unviable when less than one. BCR is expressed as:

BCR = I/C_total,

(2)

The Nepalese currency was converted to an Australian dollar value (88 Nepalese currency equivalent to $1AUD based on the typical rate observed in 2021). In this way, the net margin analysis gives the profit as an absolute difference between income and cost, with all parameters computed in a common currency (AUD). However, this absolute profit measure does not allow direct comparison between countries due to differences in purchasing power (i.e., $1AUD of net profit achieved in Australia carries a different value compared with $1AUD of net profit in Nepal). Therefore, a relative measure of costs and benefits was performed. By taking the ratio of income relative to cost, the resultant BCR is independent of the currency in which the cost and income were calculated.

The economic model of UA assumes a typical retail price can be achieved for UA produce, reflecting either marketing directly to consumers or self-consumption by household growers (displacing food that would have otherwise been purchased at retail cost). The average retail price of high-value (maximum possible value generated from the crops, for example, salad mix) and low-value crops (maximum possible retail value through the cultivation of typical vegetable crops, for example, carrots) was acquired from a review of pricing at popular supermarkets in Adelaide (Woolworths Supermarkets) [34]. Likewise, in Kathmandu Valley, the retail pricing information for high-value and low-value crop were acquired through a market information website [35]. The net margin from high-value and low-value crops under high yield, base yield and low yield scenarios were calculated using the above calculation method. For each yield scenario, three production labour input cases are considered: high (6.0 h/m²/yr), base (1.8 h/m²/yr) and low (0.3 h/m²/yr). The variation in labour input reflects how intensive the gardening/farming activity is, and (in subsequent calculations) these labour cases will map to corresponding high, base and low yield cases.

The annual gross income, I ($/yr) from UA under different yield scenarios is calculated for high yield, base yield and low yield cases using the retail market price of high-value and low-value crops by using the following formula:

I = Y × A × P_retail,

(3)

where:

Y = Yield (low, base, and high) in kg/m²/yr

A = Area (m²)

P_retail = Retail price ($/kg)

The total annualised cost C_total ($/yr) is the sum of land, labour and distribution cost:

C_total = C_lan. + C_lab. + C_dis.,

(4)

where:

C_lan.= Annual cost of land ($/yr)

C_lab.= Annual cost of labour ($/yr)

C_dis.= Annual distribution cost ($/yr)

In this study, the annualised cost C_lan. ($/yr) was approximated by evaluating 40 different parcels of land for sale and calculating mortgage repayments using a constant interest rate of 3% over a 30-year period. The cost was computed using the following formula:

$${\mathrm{C}}_{\mathrm{lan.}}=\mathrm{P}\frac{\left(\mathrm{r} {\left(1+\mathrm{r}\right)}^{\mathrm{n}}\right)}{\left( {\left(1+\mathrm{r}\right)}^{\mathrm{n}}-1 \right)}’$$

(5)

where:

P = Land price ($)

r = Interest rate (% per year)

n = Payment period (years)

The labour cost C_lab. ($/yr) was assumed to be the sum of three distinct activities: production, packaging, and distribution:

C_lab. = C_lab,pro. + C_lab.,dis. + C_lab,pak.,

(6)

where:

C_lab,pro. = Production labour cost ($/yr)

C_lab.,dis. = Distribution labour cost ($/yr)

C_lab,pak. = Packaging labour cost ($/yr)

The production labour cost C_lab,pro. was estimated based on the area under production:

C_lab,pro. = W_pro. × L_pro. × A,

(7)

where:

W_pro. = Wage rate for production labour ($/hr)

L_pro. =Labour productivity (hrs/m²/yr)

The distribution labour cost C_lab,dis. ($/yr) was estimated based on the hours spent freighting produce to customers:

C_lab,dis. = W_dis. × T × D/T_spe.,

(8)

where:

W_dis. = wage rate for distribution labour ($/hr)

$\mathrm{T}$ = Number of trips per year

$\mathrm{D}$ = Trip distance (km)

T_spe. = Average trip speed (km/hr)

For simplicity, W_dis. may be assumed to be equal to W_pro.

The packaging labour cost C_{lab, pak}. ($/yr) was estimated based on the efficiency with which the UA produce can be packed, using the following formula:

C_lab,pak. = W_pak. × Y × A,

(9)

where:

W_pak. = Payment rate ($/kg)

For calculating non-labour distribution cost C_dis. ($/yr), the estimated production was first calculated by multiplying the land area by the average UA crop yield, giving the total harvest (kg/yr). The maximum freight capacity per distribution trip (assumed to be by car) was taken to be 200 kg/trip. The total number of trips was therefore estimated through simple division and rounded up to the nearest integer. The vehicle running cost was then estimated based on a per kilometre cost, using the Australian Taxation Office rules for vehicle tax deductions as a guide [36]. The packaging material cost was estimated based on available market cost information. For this study, the distribution process was simplified to a straightforward activity of packing, driving and dropping the product from the production point to the city centre without considering the time consumed for selling the products.

C_dis. = (T × (C_v × D)) + C_mat. × Y × A,

(10)

where:

C_dis. = Non-labour distribution cost ($/yr)

T = Number of trips per year

C_v = Vehicle running cost ($/km)

D = Trip distance (km)

C_mat. = Cost of packaging ($/kg)

The number of trips per year, T, is used in both the labour and distribution cost calculations above and is given as follows:

T_. =max (52, ROUNDUP (Y × A/F)),

(11)

where:

F = Maximum freight capacity (kg/trip)

and ‘ROUNDUP’ implies rounding up to the nearest integer value.

T is taken as the maximum of either 52 (i.e., one trip per week, assumed to be the minimum frequency to take produce to consumers) or the value calculated by dividing the production quantity by the vehicle capacity (i.e., more than one trip per week).

2.4. Scenario Parameters

The list of abbreviations and values used in the formulas for calculating the results is presented in

Table 1. Descriptions and list of values used in calculations.

Symbol & Units	Description	Values		Source(s)
		Adelaide	Kathmandu Valley
Wpro. and Wdis. ($/hr)	Wage rate	25.24	0.875	[37,38]
Lhou. (hrs/m2/yr)	Low yield labour productivity	0.3	0.3	[39]
	Base yield labour productivity (non-mechanised gardening)	1.8	1.8	[28]
	High yield labour productivity (non-mechanised gardening)	6.0	6.0	[5]
Wpak. ($/kg)	Wage rate for packing	0.19	0.06	[37,38]
Y (kg/m2)	Base marketable yield scenario	5.08	1.95	[26,40]
	Low marketable yield scenario	0.24	0.24	[26]
	High marketable yield scenario (soil-based mixed UA system)	6.19	6.19	[5]
F (kg/trip)	Maximum freight capacity of car	200	200	Assuming a medium-sized passenger car, e.g., station wagon
Cv ($/km)	Vehicle running cost	0.72	0.72	[36]
Cmat. ($/kg)	Packaging material cost	0.1	0.02	Arbitrary value based on market information
Pretail($/kg)	Retail price of high-value crops (salad)	22.22	3.69	[34,35]
	Retail price of low-value crops(carrot)	2.00	1.00
$T_{Spe.}$ (km/hr)	Average trip speed	60	60	Assumed the same trip speed in both contexts

.

2.5. Land Area, Cost, Distance, Wage and Distribution Cost Information

Information on land area, cost and distance from the city centre of Adelaide, South Australia (Figure 2) were acquired from 40 sites advertised on a widely used real estate website https://www.realestate.com.au (accessed on 25 December 2020). Similarly, the land area, price and distance from city centres of Kathmandu Valley Nepal (Figure 3) were acquired from 40 advertised locations on real estate websites https://www.housingnepal.com (accessed on 26 December 2020) and https://hamrobazaar.com (accessed on 26 December 2020). Land parcel areas ranged from 154 m² to 2705 m² for Adelaide and 127.2 m² to 7949.1 m² for Kathmandu Valley. The land costs advertised through these real estate websites were used for this study. The farthest vehicle travel distances were up to 57 Km in Adelaide (see Figure 2), representing a trip from the Adelaide Central Business District (CBD) to the suburb of Gawler. In Kathmandu Valley, the trip distance (Figure 3) was considered from Central to the Suburb of Kathmandu Valley, i.e., 19 Km. These distances were taken as the assumed trip distance for produce distribution.

The local wage rate information of Adelaide and Kathmandu Valley were collected from Safe Work SA [37] and the Himalayan Times [38]. The information on distribution costs is derived through the potential yield from the available area and distribution distances.

For the purpose of this study, we categorised distances and areas into simplified groups: inner-city (up to 10 kilometres from CBD), Suburban (greater than 10 kilometres from CBD), smaller-plot (up to 400 m²) and larger-plot (greater than 400 m²). Then, we calculated cost and production under different scenarios based on the land, labour and distribution cost information from the above areas and distances.

3. Results

3.1. Base Land, Labour, and Distribution Cost Analysis

Based on the available information of 40 different parcels and distances of land collected, the distance and area-wise cost (total land, base production labour, and distribution) analysis between Kathmandu Valley and Adelaide were conducted with results presented in Figure 4A,B.

The distance-wise analysis under base case production (Figure 4A) showed that land cost dominates the overall cost in both cities for inner-city UA (within 10 km of CBD). The average price of inner-city UA land in Kathmandu Valley was somewhat higher than Adelaide ($107.7 versus $90.43/m²/yr), with higher variability of price in Adelaide compared to Kathmandu Valley. The average total cost was more in Adelaide ($143.42/m²/yr) than in Kathmandu Valley ($110.58/m²/yr) due to Adelaide’s much higher labour costs ($51.11/m²/yr). For Suburban UA, the cost of land dropped to approximately one-third of the inner-city case, with the analysis again showing slightly higher average land costs for Kathmandu Valley ($38.59/m²/yr) compared with Adelaide ($33.53/m²/yr) but this time with much higher variability in land cost in Kathmandu Valley. However, the average total cost for Suburban UA was almost two times higher in Adelaide than in Kathmandu Valley ($92.92/m²/yr versus $45.22/m²/yr) due to Adelaide’s higher labour costs. Non-labour distribution costs were minimal.

In areawise analysis (Figure 4B), the average total cost for smaller-plot UA in Kathmandu Valley was almost equal to Adelaide ($112.65 versus $115.27/m²/yr, respectively). The labour and distribution costs were a negligible component of costs in Kathmandu Valley, but due to nearly two-fold higher land prices ($108.76/m²/yr) in Kathmandu Valley, the average total costs in both contexts were almost identical, and the land price variability is almost the same in small plot UA in both contexts. Interestingly, however, the larger-plot UA showed a higher total cost in Adelaide ($94.49/m²/yr) than in Kathmandu Valley ($52.62/m²/yr) despite high land price variability in Kathmandu Valley, and much higher average land prices in Kathmandu Valley ($50.32/m²/yr) compared to Adelaide ($96.86/m²/yr).

3.2. Non-Zero and Zero-Cost Analysis

Several scenarios were considered to evaluate the conventional economic viability of UA based on plausible combinations of land, labour and distribution costs. Two land scenarios were considered: full cost (representing ‘vacant lot UA’ where market cost must be paid for the lots of lands), and zero land (representing ‘backyard/rooftop UA’ where growing space is accessed free of charge). For each land scenario, three production labour input cases are considered: base, high and low.In addition to these six land/labour combinations, a seventh scenario was considered where land was costed but production labour was free (representing ‘discretionary/hobby UA’ where gardening is considered to be done in spare time with negligible opportunity cost).

For brevity, alternative distribution cost scenarios are not investigated here due to the minimal contribution of these costs to the total; distribution costs are therefore calculated as per the previous analyses. The same land parcels have been used in the subsequent analysis to provide a comparable range of UA plots and ensure a comparable range of land areas and distribution distances as in the baseline analysis (recognising that the actual farming in the backyard/rooftop scenario would necessarily take place on different parcels of land than these vacant lots).

The vacant lot UA cost analysis under high and low labour cost scenarios is presented in Figure 5A,B. The distance-wise analysis under the high labour case (Figure 5A) showed that labour cost dominates the overall cost in Adelaide while land cost is the prime factor in Kathmandu Valley; in the low labour cost case, land cost dominates in both cities. The average cost of inner-city UA under the high case is more than two times higher in Adelaide compared to Kathmandu Valley ($250.40 versus $114.59/m²/yr), while in the low case, the total costs are slightly higher in Kathmandu Valley. For Suburban UA, the cost of land dropped to approximately one-third of the inner-city case for both cities, with the analysis again showing slightly higher average land costs for Kathmandu Valley with much higher variability. However, the average total cost for Suburban UA was almost 4.6 times higher in Adelaide under the high labour cost case ($199.92 versus $49.30/m²/yr), while for the low case, the cost is slightly higher in Kathmandu Valley due to the high sensitivity to labour costs in Adelaide and the dominance of land costs in Kathmandu Valley.

In area-wise analysis (Figure 5B), the average total cost under the high labour case for smaller-plot UA in Adelaide was nearly two times higher than in Kathmandu Valley ($222.25 versus $116.65/m²/yr, respectively), while for the low case, the total cost seems higher in Kathmandu Valley due to much higher land costs ($108.76/m²/yr). In both contexts, the land price variability is almost the same in smaller plot UA. The larger-plot UA showed a higher total cost in Adelaide ($201.53/m²/yr) than in Kathmandu Valley ($56.68/m²/yr), while in the low case, the price was almost the same in Kathmandu Valley ($51.16/m²/yr) and Adelaide ($51.46/m²/yr). The costs seem to be largely driven by labour costs in Adelaide and land costs in Kathmandu Valley. The non-labour distribution cost remains a very small fraction of overall cost in all scenarios.

The per square meter cost analysis under zero land price with different labour scenarios (backyard/rooftop UA) under base, high and low input is shown in Figure 6A,B. The distance-wise analysis under base case showed that labour cost dominates the overall cost in Adelaide and Kathmandu Valley. The average cost of inner-city UA is significantly higher in Adelaide due to Adelaide’s 28 to 33-fold higher labour cost. The same result was observed for Suburban UA. The average total cost for Suburban UA was almost 15.5 times higher in Adelaide under the high case ($166.38 versus $10.71/m²/yr), while for the low case, the cost was over 11 times higher in Adelaide due to the dominance of labour cost in Adelaide.

The average cost under the high and low cases for smaller-plot UA in Adelaide is much higher (14.8, 20.8 and 6 times, respectively) than in Kathmandu Valley (see Figure 6A,B). The larger-plot UA showed much higher price variation in Adelaide (23.7, 25.14 and 13.7 times, respectively). The costs seem to be largely driven by higher labour costs in Adelaide. The non-labour distribution cost appears to be insignificant in all scenarios.

The cost of UA under the non-zero land and distribution (discretionary labour/hobby UA) in Adelaide and Kathmandu Valley (Figure 7A,B and Figure 8A,B) showed the dominance of land cost in all cases, as the land cost is now a much higher fraction of the total cost. In Suburban and inner-city UA occurring in Adelaide (Figure 7A and Figure 8A), the proportion of land costs represented in the total costs under the base, high and low cases was much higher than in the inner city (97 to 98 % in the inner city and 84 to 86% in Suburban). The same phenomenon was observed in the Kathmandu Valley (see Figure 7A and Figure 8A) case (98% in inner-city and 88% in suburban). The average total cost of Suburban and inner-city UAs was slightly higher in Kathmandu Valley than in Adelaide due to the dominance of land cost. The higher variation in inner-city UA total cost under three cases was observed in Adelaide, while in the Suburban case, the variation was greater in Kathmandu Valley.

Likewise, in smaller plots, the land cost was greater than distribution costs, with 92% share of total cost in Adelaide and nearly 93% in larger plot UA (Figure 7B and Figure 8B). Similarly, in Kathmandu Valley, the land represents a high proportion of cost (98% in smaller plots and 93% in larger plots). In the smaller plot and larger plot UA (see Figure 7B and Figure 8B), the total cost share of Kathmandu Valley was found to be higher in all cases, with many variations in land price in Adelaide and almost the same variation of total cost between Adelaide and Kathmandu Valley under both areas.

3.3. Gross Margin Analysis

The gross margin was calculated for the 40 parcels of land in Adelaide and Kathmandu Valley. The three yield scenarios under two crop combinations were considered: high yields and high-value crops (representing maximum possible income under a combination of high retail value crops), high yields and low-value crops (maximum possible income under low retail value crops), normal yield and high-value crops (‘typical’ income under a combination of higher-value crops), normal yield and low-value crops (‘typical’ income under lower-value crops), low yield and high-value crops (lower-range estimate of income based on a combination of low yield but high-value crops) and low yield land low-value crops (lowest possible income under the unfavourable combination of low-value crops and low yields).

The per square meter gross margins from UA under three different yields and two crop scenarios are presented in Figure 9. The difference in crop prices between the two cities is very large, with retail prices in Adelaide much higher than in Kathmandu Valley. This is in contrast to the total costs calculated earlier, whereby in some cases, Kathmandu Valley was just as costly—or even more so—compared with Adelaide. So, counterintuitively, economic viability may be more achievable in Adelaide (due to high retail prices) than in Kathmandu Valley, at least when land costs are included. This analysis shows that crop choice is critical: even with high yields, the gross margin from low-value crops is small.

3.4. Net Margin Analysis

NM is calculated for both high- and low-value crops, giving a total of 18 combinations each for Adelaide and Kathmandu Valley. The margins from three UA types calculated under different UA scenarios are presented in Figure 10A,B, Figure 11A,B and Figure 12A,B.

The distance-wise margin analysis under vacant lot UA in Adelaide (Figure 10A) showed negative results irrespective of crops and production scenario in all cases except the Suburban base case under high-value crops ($19.96/m²/yr). The area-wise analysis (Figure 10B) showed the same, negative margin in a smaller plot (although some variability suggests a positive margin may sometimes be achievable). NM was only positive for high-value crops under the base case growing in larger plots (18.39/m²/yr). In Adelaide, a high-value crop under base case production labour was the only positive margin for vacant lot UA, suggesting that the costs of additional labour were not offset by the additional yields. The other vacant lot UA options showed negative margins across Adelaide’s entire range of land parcels. In Kathmandu Valley, the average margin was negative irrespective of area and distances in all vacant lot UA scenarios (see Figure 10A,B), albeit with greater variability than in Adelaide.

In the NM analysis for backyard/rooftop UA in Adelaide (Figure 11A,B), the minor variability in results reflects the variability in distribution costs that arise due to variations in parcel size and travel distance. The distance and area-wise analysis showed positive results in all distances and areas under the base yield case growing high-value crops, with the highest margins in inner-city followed by larger plots ($59.89/m²/yr and $58.25/m²/yr, respectively). Likewise, the high-value crops under base case in smaller plots and Suburban showed encouraging margins in Adelaide.

In Kathmandu Valley (see Figure 11A,B), positive margins were found in the inner-city, Suburban, smaller and larger plots under high-value crops and both high labour and base production scenarios A better margin was found in the case of high-value crops with high yield (and corresponding high labour) in the larger plot ($16.23/m²/yr) followed by smaller-plot ($14.95/m²/yr). Likewise, the high-value crops under base yield in larger plots and the inner city showed positive margins ($4.89 and $4.32/m²/yr, respectively).

In Adelaide, discretionary UA showed a potentially viable result (on average) for high-value crops under both high yield and base yield scenarios within different distances and scales (Figure 12A,B). Better margins were found in high-value crops under high yield in Suburban ($97.98/m²/yr), followed by high-value crops with high yields from larger plots ($94.62/m²/yr). Unlike the previous scenarios for Adelaide, when labour is free, it becomes beneficial to pursue the highest yields. In contrast, in the Kathmandu Valley case (see Figure 12A,B), all scenarios showed negative margins on average, with substantial variability. Kathmandu Valley’s results for this version of hobby UA were similar to those for vacant lot UA due to the dominance of land cost over labour. Despite low labour and distribution costs, the high land cost and low retail price of vegetables may be the prominent causes of low margins in Kathmandu Valley compared to Adelaide.

3.5. Benefit–cost Ratio (BCR) Analysis

By definition, this analysis identifies a positively viable situation (BCR > 1) in any case with a positive NM, such as Adelaide’s Suburban and larger plot cases. The BCR analysis becomes most useful when we compare scenarios that show positive NM in both cities. For backyard/rooftop UA in both contexts under two distances and areas (Figure 13A,B), there is positive viability under high-value crops grown under base yield and labour in Adelaide. In Kathmandu Valley, high-value crops grown under either base yield or high yield conditions showed BCR > 1, albeit much wider variability in larger plots. Comparing the viability of UA between the two cities using BCR (which is independent of currency) gives a different impression from that obtained using the absolute NM. In the potentially viable cases, the NM shows a much higher absolute margin may be in Adelaide than in Kathmandu Valley; however, comparing the BCR values suggests that in Kathmandu Valley, the income represents a much larger amount relative to the cost, and hence viable UA in Kathmandu Valley may deliver approximately twice the return (relative to the input) compared with viable UA in Adelaide.

4. Discussion

4.1. Urban Agriculture Cost and Margins

It is noted that repeating the analysis for specific crop types, specific locations and taking into account seasonal production opportunities would allow these results to be further refined, especially for individual cases.

The land costs are likely to be prohibitively expensive for commercial UA to be profitable, especially in the inner city and small-plot cases, but if UA practitioners can access land at the lower end of the price range, then there is a higher chance that it may be economically viable. Another key finding is that labour costs in a high-income setting like that in Adelaide are similar to land costs, whereas in Kathmandu Valley, labour costs are a small fraction of overall costs, and the total cost is overwhelmingly driven by land costs.

The cost analysis showed wide variations between Adelaide and Kathmandu Valley due to very large labour cost differences. Australia’s labour cost is the third-highest globally [41].

Therefore, it is only by choosing high-value crops that farmers can potentially generate profit in Adelaide, and this modelling suggests that profitability may only occur under the base yield scenario, as the additional cost of labour to achieve high yield would outweigh the extra income. While these results reflect the assumed values used in the analysis, they point to the need to carefully consider the marginal gain from investment in extra labour in pursuit of higher yields. In Kathmandu Valley, if expensive land costs can be avoided, growers can make a small profit through high-value crops in both base and high-yield conditions. Furthermore, the BCR analysis suggests that these modest profits in Kathmandu Valley may in fact being a greater benefit (relative to input cost) than the equivalent cases in Adelaide that carry a larger absolute profit.

In Adelaide, the focus should be given on high-input farming for the high yield and income without the additional cost offsetting the income, and the high retail price of products may overcome the high land cost in some cases. Meanwhile, in Kathmandu Valley, the labour cost is already low compared to the land cost, so making labour free does not substantially affect the outcome: if the growers have to pay the full cost of land, UA will be unviable.

Under the ‘vacant lot’ scenario, growers can only generate profits in locations similar to Adelaide by growing high-value crops, like salad mix. If paying full land costs, the growers will always make a loss in Kathmandu Valley as the average land price of Kathmandu Valley is almost two times higher than Adelaide, while the retail price of UA products is comparatively low. Growers will make a net loss in both developmental contexts under vacant lot UA from low-value crops. If land can be accessed freely, UA is far more likely to provide a viable return on labour when practiced in Kathmandu Valley, while there is potential to deliver a marginal return on labour when practiced in Adelaide, ideally also with free land access.

For instance, comparing the result from 51 studies, most American and Canadian experts expressed viability based on measuring the extent to which the income needs of the farm family are met while in contrast, European experts described viability as an opportunity cost of capital investment [15]. There may be a possibility of getting maximum margin for viable UA by cultivating high-value crops like herbs and salads but selecting a cost-effective and scalable UA crop is important for maximising profit [12].

4.2. The Scale of Operation, Distribution Distance and Viability of UA

Despite the enormous potential market, commercial UA farms are still uncommon in most cities due to factors like expensive land and limited plot size availability [21]. The current analysis has quantified the magnitude of the potential costs and income from UA, demonstrating the challenge faced by farmers seeking to make a commercially viable enterprise.

Previous researchers [22] found high initial investment, expensive land, and labour cost as the leading causes of high food costs in the cities. Other factors beyond the cost of production may be associated with the economic viability of urban farms, like demand from the farmers’ market and restaurants [23]. This is important to note because the economic model in this study assumes that all produce was sold, and all results are compromised should a fraction of produce remain unsold or wasted due to limited demand. The results are also dependent on zero cost for market access—further refinement to the model may be made to consider the cost of accessing a farmers’ market or similar environment to sell produce for commercial UA scenarios. The high land cost in both contexts and high labour cost in Adelaide were major contributing factors to increased production costs. Due to high associated costs during production and distribution, urban growers must think about producing high-value specialty products for a niche market to maximise profitability [16]. However, even in high-value crops, the price of the respective high-value commodity relative to the associated costs of land and/or labour will decide the level of economic viability. It is important to note in this context that external factors influence the viability of UA, including the terms adopted in this model. For example, changes to minimum wage rates and supply chain interruptions which increase sale prices can positively and negatively impact viability.

Interestingly, smaller UA gardens have been shown to carry an economic advantage over larger ones by allowing the intensification of high-value crops [12]. Generally, a high margin is driven by the retail price rather than other production inputs and in Adelaide’s conditions, the net value is higher in small-sized gardens than in larger ones due to the selection of high-value crops [12]. The current analysis has shown that growers can potentially achieve a net profit by choosing high-value crops in both contexts, if land cost is minimised, which may require subsidisation through policy mechanisms as cost benefit streams are the key to the financial viability of UA [42]. Growers should then consider increasing yield to improve income. However, if higher yields are achieved through more labour-intensive production processes, then depending on the development context (which influences wage rates and retail prices for produce grown) there is likely to be an optimum level of intensification where the increased labour costs are compensated by increased income, and further increases in labour inputs may be uneconomic.

The problem of land scarcity and high cost may be somewhat mitigated through vertical or rooftop farms, but these options also require some investment in setup and operational cost. As such, it is plausible that the food produced in these settings may still need to be high-value niche products, which may remain inaccessible to mid-to low-income people in urban areas. For this reason, backyard or community gardens are the most practical and low-cost options and UA is indeed practiced as a subsistence activity by many city dwellers [4]. Even where cultivation takes place in small areas in subsistence conditions, it can provide an economic benefit by saving the amount of food that a household must purchase from markets [43].

Specific UA setups may give higher productivity, lower labour input, and/or lower costs than others [29], so there are opportunities for growers to think about alternative low-cost UA that may improve profitability beyond the parameters considered in the current analysis. Among the 34 studied edible food gardens of South Australia [28], 65% would be at break-even within five or fewer years without considering land or labour costs, and one in five gardens could be financially viable with minimum wages applied to labour (but still zero land cost). Interestingly, in that study, the majority of the South Australian home gardeners were found to be practising UA—at least in part—to save money, and the majority believed that they were somehow successful in that goal [28]. The present study has pointed to the scant chance of financial viability (which can be taken as a proxy for saving money) when considering non-zero land, labour, and distribution costs under normal yield conditions in South Australia. This suggests that if UA is practised in pursuit of monetary savings, the savings must effectively be subsidised through discounted (or free) access to land, and/or through the use of discretionary (free) labour.

The current analysis suggests that high yield resulting from high labour inputs will improve the viability in Kathmandu Valley, where labour costs are low relative to retail food prices, but the same strategy may be unviable in Adelaide, where the increased labour costs may outstrip the increased income. The analysis in this study has demonstrated the interplay between land and labour costs, yield and labour input, and retail prices that collectively govern the financial viability of UA as a business model.

The BCR analysis shows the viability of backyard/rooftop UA in Adelaide and Kathmandu Valley under high-value crops under the base and high yield scenarios in two UA distances and scales. The discretionary labour UA in Adelaide also showed chances of viability under a high-value crops base and high yield in the given distances and scales and vacant lot UA under Suburban and larger plots. Urban Agriculture exists as traditional soil-based farming in home gardens and available vacant areas [22] in most parts of the world. The high initial investment, expensive land and limited plot size have already been identified as economic issues in UA, which can be minimised through the production of niche products, adoption of business models and finding inexpensive land plots [21]. UA’s financial viability can be further increased by choosing short cycle and high-value crops grown in unique structures, and cost can be reduced through the use of community resources like volunteer labour and low-cost land lease agreements with the local government [44]. Hi-tech UA, including vertical farming, are emerging practices in UA nowadays, but they are costlier due to initial investment in infrastructures and additional operational costs. Thus Backyard gardens and community gardens may be crucial alternatives to reduce the cost of urban foods at a cheaper price than hi-tech UA [4]. To be sustained financially, UA farms could also focus on activities beyond production, like agrotourism, training, consumer workshops, and events [23].

4.3. Policy Implications

It has been recommended that a better policy environment is required to improve access to land and financial support for better productivity and profitability of UA [17]. Others have pointed to increasing land prices as the major threat to accessing suitable land for better productivity of UA, which is broadly supported by the present study [23]. Compared to non-agricultural land use, the value generated from UA to the city may not be large [45], but it may add value indirectly, e.g., by increasing the surrounding property value [46]. UA developers and planners should consider the broader role of UA—beyond the economic viability of individual food production endeavours—in terms of increasing local food production capacity, reducing food miles and employment generation [47].

To produce local food sustainably, there have been calls made for a better food plan for cities; this involves considering the total costs for UA production [28]. Based on the findings from this study, the following policy measures are recommended for the financial viability of UA in the long run:

• The UA practices seem unrealistic under full market cost consideration for land. So, irrespective of the development context, governments should consider free access to public land, vacant land, or subsidised land for UA. Others have recommended a suitable land conversion mechanism as one of the essential policy tools for availing cheap land for UA activities [48].
• Labour input is directly related to UA productivity, and the cost of labour in high-income cities (e.g., Adelaide) is significantly higher than in low-income cities (e.g., Kathmandu Valley). This analysis has highlighted the importance of considering the wage rate, the production input (hr/m²) and the labour-yield relationship all within the context of retail prices, to identify an optimum level of labour input. Appropriate labour-saving technology (e.g., modest scale mechanisation) may be a key part of delivering viability in high-wage scenarios that supports UA practitioners in planning and accessing appropriate technologies to optimise labour input and yield/income.

5. Conclusions and Recommendations

The purpose of this investigation is to subject UA scenarios to a basic cost/income analysis to identify the potential range of cases in which outright economic viability may be achievable and elucidate the most effective policy actions to improve viability. The analysis has shown that the economic viability of UA depends on the profitability generated through UA practices, and the choice of crops primarily determines whether a viable level of income will be attained. Therefore, a careful choice of commodities is important for the sustainable uptake of UA across cities.

However, the income from crops is only one-half of the viability equation. The sustainability and viability of UA depend critically upon the cost of inputs, especially land and labour. This study demonstrated a method for economic quantification of land, labour and distribution costs for checking the economic viability of UA. Applying the cost analysis in Adelaide (high-income) and Kathmandu Valley (low-income) showed little chance of economic viability under the ‘vacant lot’ scenario (full land, labour and distribution costs). However, in Adelaide, there is a chance of viable vacant lot UA by selecting high-value crops in Suburban areas (with lower land costs than the inner city) due to the premium market price for produce. Vacant lot UA seems unviable in Kathmandu Valley due to the low market prices, even for high-value crops, due to high land prices.

In cases where land cost is eliminated, both Adelaide and Kathmandu Valley show a greater likelihood of achieving viability, but the balance between the amount of labour input and the resultant yield is important; in Adelaide, the highest input/highest yield case is not the most profitable, while in Kathmandu Valley the low labour costs relative to retail prices favour investing more labour. These results point to the importance of considering labour productivity, and there is scope for future research into appropriate labour-saving technology.

Interestingly, using a BCR analysis (ratio of income to cost), the viable cases in Kathmandu Valley represent a proportionally greater benefit than the corresponding viable cases in Adelaide, despite the net margins being substantially higher (in absolute terms) in Adelaide. This suggests that local conditions—especially the value of production labour relative to the value of food produced—are critically important when considering the economics of UA.

The method presented here has some limitations. The example land parcels covered 40 different distance/area combinations for both Adelaide and Kathmandu Valley but may not represent the overall scenario globally. Likewise, the study excludes the potential cost and contribution from high-tech UA practices including vertical and indoor farms, other impacts like taxation, subsidies available in local markets and does not consider other costs which larger operations might consider like risk (in terms of occasionally failed crops), water for irrigation and insurance against unforeseen impacts on the production. The retail costs were all derived from a large supermarket, access to these markets also requires high costs in quality management and control and as such, urban gardens are at present most likely to be restricted to community markets.

However, the results help shed light on key considerations for UA as a sustainable business model. Ideally, UA activities would represent a local food production system that can make a reasonable contribution to food security by delivering food at a cost that is competitive with typical market prices. To this end, land and labour costs are identified as key barriers to the financial viability of UA, and thus urban planning processes could facilitate cheaper land, and subsidised labour arrangements and/or appropriate labour-saving technology could lead to economic advantages. It is also recommended that the economic modelling presented in this study be refined by the collection of better data, particularly in the volume of production by UA, especially for high-value crops which are more likely to have commercial profitability, the costs of accessing a market and critically, the volume of product in commercial and domestic settings that remains unsold or unused, which can impact on commercial viability. The practice of backyard/rooftop UA in the context of low-income cities like Kathmandu valley, where the land cost is potentially high and discretionary labour UA with modest mechanisation in the high-income setting of Adelaide, Australia, where labour cost is the primary threat, can improve economic viability.