An economic approach 
to tackling antimicrobial 
resistance (AMR) 

A thought provoker 

1 

Author 

Professor Lin Crase 
Professor of Resource Economics, UniSA Business 
University of South Australia 

Editorial contributions from 
Dr Branwen Morgan and Chris Angwin 
Minimising Antimicrobial Resistance Mission 
CSIRO 

Adapted from work by 

Dr Bethany Cooper 
Senior Research Fellow, UniSA Business 
University of South Australia 
bethany.cooper@unisa.edu.au 

Professor Harry Clarke 
Emeritus Professor, Business School 
La Trobe University 

Professor Lin Crase 
Professor of Resource Economics, UniSA Business 
University of South Australia 

Dr Ben van den Akker 
Lead Wastewater Scientist 
SA Water 

Professor Erica Donner 
Research Director, CRC SAAFE 
University of South Australia 

Dr Branwen Morgan 
Minimising Antimicrobial Resistance Mission 
CSIRO 

Date of publication 

August 2023 


The problem 

Antimicrobial drugs work by destroying or hindering 
the harmful microbes that cause infections (pathogens). 
However, microbes can rapidly evolve to resist the 
effects of medicines designed to kill them. 

Therefore, over time, antimicrobials become less 
effective. This process is accelerated by the use and 
misuse of antimicrobials. 

This phenomenon, known as antimicrobial resistance 
(AMR), is a major international concern. 

Although fears about the potential disruption from 
AMR have existed for decades, mounting evidence of 
its impacts on human health has elevated its status to 
a “global health threat”1. 

In 2014, global deaths from AMR were estimated at 
around 700,0002. However, by 2019, the death rate 
had risen significantly to 1.27 million per year3. 

By 2050 this is expected to increase to 50 million 
deaths per year2 and cause an increase in global AMR 
healthcare costs from $300 billion to greater than 
$1 trillion per annum4. 


1 World Health Organization (WHO) (2015) Global Action Plan on Antimicrobial Resistance, available at www.who.int/publications/i/item/9789241509763 
2 O’Neill J. (2014) Tackling drug-resistant infections globally – Final report and recommendations, The Review of Antimicrobial Resistance, amr-review.org/. 
3 Antimicrobial Resistance Collaborators (2022) Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis, Lancet; 399: 629–55 
4 WHO. 2019. “Ten threats to global health in 2019”, Newsroom, www.who.int/news-room/spotlight/ten-threats-to-global-health-in-2019 


One Health 

Initially, concerns about AMR were focused on human 
misuse of antimicrobials. In 1945, Alexander Fleming 
himself observed that the antibiotic he discovered, 
penicillin, could lose effectiveness due to the 
evolutionary process of antimicrobial resistance5. 

Now, we understand AMR in a broader context, 
considering humans alongside animals, food 
production and the environment. 

There has been a strong focus on use of antibiotics 
in animal production particularly, building directly on 
the links between animal and human health, dating 
from the 19th Century. 

Subsequently, emphasis has shifted to include wider 

environmental influences and the notion of ‘One 

Health’ has been used to characterise the complexity 
of the challenge. 

One Health is premised on the idea that human, 
animal, and environmental health are interlinked and 
that actions to reduce AMR need to be coordinated 
across these domains. 

The least-well understood component of the One 
Health perspective on AMR is the role that the 
environment plays in eventual clinical presentations. 

Despite very few antimicrobial active pharmaceutical 
ingredients (APIs) being manufactured in Australia, 
the release of unmetabolised or partially metabolised 
antimicrobials into water and the environment is 
acknowledged as a driving force of AMR. 

Some antimicrobials are packaged in Australia and 
many are delivered/formulated in animal feed. 

Currently, there are no Australian regulations for 
pharmaceutical levels in waste discharges (hospital 
or household). 

The only Australian ‘regulation’ or guideline that 

mentions antimicrobials is the Water Quality 
Guidelines for recycled drinking water 2008. 
This has default guidance values for antimicrobials 
and chemicals. 

One 
HealthAnimalhealth 
Environmentalhealth 
Human health 
The problem is that pharmaceuticals (unlike 
chemicals) are not included in most regulations and 
guidelines related to waste in Australia and globally. 

Furthermore, specific guidelines or regulations on 

hospital waste/wastewater treatment plants do not 
consider antimicrobials. 

Regulations dealing with medical waste (infectious, 
hazardous, radioactive, and general) do not cover 
drugs/antimicrobials entering the sewage waste 
system (urine/faeces). The only reference to 
pharmaceuticals is under hazardous waste disposal of 
unused medications. 

Globally, there is an increased focus on the removal 
of pharmaceutical residues from wastewater. In the 

Netherlands, the government adopted the ‘Chain 

Approach Medicine Residues from Water’ policy. 

The chain approach has recently been nominated 

by the ‘World Future Council’ for the Future Policy 

Award 2021. The Dutch Consortium on Antibiotics 
and Pharmaceutical Residues from Water consisting 

of a ‘coalition of doing’ acts in accordance with this 
chain approach specifically focuses on the removal of 

antibiotic residues from water. 

5 Spellberg B, Gilbert DN. (2014) The future of antibiotics and resistance: a tribute to a career of leadership by John Bartlett. Clin Infect Dis. 59 


Why different perspectives are 
needed 

While tackling AMR through wastewater has clear 
merit on some fronts, the complexity and unknown 

aspects of AMR make it difficult to identify 
definitive solutions. 

Accordingly, bringing different perspectives to bear 
when contemplating actions is helpful. This is a core 
principle embedded in the One Health approach. 

In addition to recognising that humans are one aspect 
of the ecosystem, it is important to also recognise that 

human behaviours in that environment are not fixed, 

and that can result in feedback effects, both positive 
and negative. 

For instance, the price of alternatives to antimicrobials 
used in some settings (e.g. vaccination) could change, 
resulting in less (or more) use of antimicrobials. 

International travel could become more accessible 
and affordable, accelerating the spread of AMR across 
countries. In these circumstances, understanding 
social sciences, like economics, offers an alternative 
lens for formulating solutions. 

Economic approaches to AMR 

Economics has several appealing aspects that can 
help decision-making. First, most economic analysis 
is concerned with change at the margin. That is, 
choices are considered and compared based on the 

net benefits of taking one action over another. This 

feature has resulted in a widely-used approach known 
as cost-benefit analysis (see case analysis on page 8). 

Cost-benefit analysis usually follows four rudimentary 
steps: (1) establish the scope of benefits and costs that 

attend a choice – ideally keeping the scope similar for 

both; (2) measure benefits and costs with the same 
metric, usually dollars; (3) discount future benefits 

and costs so that all values are expressed in current 
terms; (4) apply a decision rule, for instance choosing 

the alternative that delivers greatest net benefits. 

While appearing straightforward the application 

of cost-benefit analysis to specific choices can be 

challenging, particularly when information 
is incomplete. 

The second appealing aspect of economics is that 
it has established mechanisms for dealing with 
choices when information is not fully available. Here, 
economics makes a distinction between the idea of 
‘risk’ and ‘uncertainty’. Risk in economics relates to a 
decision that has a probabilistic outcome. 

5 


In contrast, uncertainty pertains to situations where 
the gaps in knowledge are so severe that a probability 
distribution is not available. In practice, many choices 
sit along this spectrum and expert judgement or 
community preference are used to shape the choice. 

Where a probability is assigned to an outcome the 

standard steps in a cost-benefit analysis can still 

be applied by invoking expected utility theory. For 
instance, if a project has an 80 per cent chance of 

delivering a benefit of $100 million dollars, this is 

weighted at $80 million and included in the cost


benefit framework. 


However, once this type of calculation enters 
the analysis it is important to also consider risk 
preferences. In the previous calculation we derive 
the $80 million by simply multiplying the value of 

the benefit ($100m) by the probability (0.8) to derive 

expected value – this implicitly assumes risk neutrality 
on the part of the decision maker. 

If the decision maker is risk averse, they will impose a 

subjective judgement that modifies the probability so 

the expected outcome will be less than $80 million. 

Risk aversion has potentially important consequences 
for the way we approach decisions that relate to AMR. 
Given the potentially catastrophic consequences of 
AMR a decision maker might see value in applying 
some for of insurance premium (see case analysis on 

page 10). within the cost-benefit analysis. 

In this case, a project might not satisfy the standard 

criteria of generating benefits in excess of costs, but 

the shortfall is less than the risk premium the decision 
maker would have paid to avoid the potentially bad 
outcomes. Clearly, the challenge with this approach is 
that an extremely risk averse decision maker accepts 
nearly all projects, even if the probability of a poor 
outcome is very low. 

Where a choice is circumscribed by very little 
information, there may also be value in holding off 
making any choice, until more information becomes 
available. Economists use the notion of real options to 
explore these types of cases. 

The idea behind real options theory is that a choice 
can sometimes be irreversible and if later proven 
wrong because additional information emerges, it is 
better to wait so that the unrecoverable investments 
are 
not wasted. 



In simple terms, the real option value acts against the 
insurance premium, described above, and requires 

benefits to exceed costs by a sufficient margin to 

offset any potential gains from waiting. The practical 
way this is often included in decision making, is 
to stage projects in smaller phases to assemble 
additional knowledge as the project rolls out, even if 
the staged approach is more expensive in standard 
cost terms. 

Applying these concepts to the removal of 
antibiotic residues in wastewater streams yields 
some interesting results. First, the choice between 
upgrading centralised wastewater treatment or more 
decentralised improvements implies some form of 
binary choice. 

Once we account for the different types of uncertainty 
this is unlikely to be the case; even a risk-neutral 
decision maker might reasonably opt to trial both, 
in order to better understand the effectiveness of 
each treatment. Second, a decentralised approach 
also carries with it the likelihood of some form of 
regulation of the decentralised entities. This has 
the attractive feature of being relatively reversible 
in an economic sense, thus avoiding large capital 
costs that might be ineffective in terms of AMR but 
nonrecoverable in centralised systems. Nonetheless, 
there are material costs related to monitoring to bring 
compliance (see case analysis on page 11) and these 
also need to be considered. 

Conclusion and next steps 

AMR is a serious risk that cannot be addressed with a 
single intervention: there is no silver bullet. The One 
Health approach helps conceptualise the challenges 
while highlighting the problem’s wicked nature. There 
is a role for the water and wastewater sector and, the 
use of economics, offers a way of considering how the 
sector might best respond. 

An opportunity exists now to engage on this issue. 
We welcome feedback on the ideas expressed in 
this thought-provoking piece and plan to release 
additional resources to support engagement. 

7 


Case analysis: 

Wastewater upgrades and cost-benefit analysis 

Contemplate a case where a wastewater utility is 
considering upgrading infrastructure to produce 
tertiary treated wastewater and subsequently onselling 
this to irrigation customers. If the utility was 

operating solely in the private sector, the cost-benefit 

calculus would reduce to estimating the extra capital 
and operating costs of the plant and distribution 
infrastructure and forecasting whether the additional 
revenues would then adequately cover the costs of 
the project. 

Here the scope of the project is defined by the 

commercial interests of the wastewater entity and 

the measurement of benefits is based on expected 

payments from customers, with some degree of 

probability to reflect the likelihood that future 

customers may pay less (or more). Construction cost 
estimates would be taken from similar engineering 
works as would operating costs. Importantly, because 
this private entity could invest in other projects, a 
cost of capital (e.g. interest foregone) would also 
need to be included in the cost calculation. 

If the wastewater company expects the project to 
last for 20 years and collect revenues over that time, 
then future revenues would need to be discounted, 
since a dollar earned in the future is worth less than 
a dollar generated today. Finally, the company would 

make a decision by considering if the benefits (in this 

case the sum of discounted expected revenues) were 

sufficiently in excess of costs to warrant any risks 

related to the project. 

While seemingly straightforward, even this relatively 
simple private sector example has a number of 

‘unknowns’. Will construction costs change? If 
energy prices shift, how will the project fare? Will 
irrigators want the water at the prices on offer? Are 

there alternative water sources the irrigators might 

be offered? Will the cost of capital change? It is 

thus common practice to use expected values but 
to then test different assumptions and scenarios 
to establish how sensitive the results might be to 

specific assumptions. 



In settings where external benefits and costs are 

major issues it makes sense for these types of 
wastewater projects to be government-owned. 
Managing wastewater has numerous health and 

environmental spill-over benefits and for that reason 

it is usually undertaken by an arm of government. 

Government ownership has implications for 

how the steps in the cost-benefit analysis above 

transpire. First, the scope of the project would 

likely change, inasmuch as a wider set of benefits 

and costs, like environmental and social impacts, 
might legitimately be of concern to a publiclyowned 
utility. That is not to say that the commercial 
aspects are ignored; rather, the project would now 
be more likely to take account of improvements (or 
deterioration) in the natural environment. 

For example, reverse osmosis might be required 
to treat AMR risks, but this produces wider 

environmental quality benefits that might now lay 

in scope. Government ownership also implies that 
additional consideration might be given to secondround 
employment impacts1. Second, this wider 

project scope means that measurement of benefits 

and costs will need to use other approaches, like 
stated-preference techniques or related-market 
techniques to determine values. Third, the discount 
rate chosen might not be the same for a 
government project. 

For example, concerns about the possible downplay 
of impacts on future generations might see a lower 
discount rate included2. Finally, the decision rule 

will be similar, although in this case the cost-benefit 

analysis will form only one piece of information 
used by the decision maker. 

Much of the information used in a cost-benefit 

analysis for a project like this would also be 
employed by any economic regulator. Centralised 
wastewater systems lend themselves to monopoly 
service provision and Australian jurisdictions 
attempt to control the behaviour of monopolies, 
including those owned by government. 

The role of an economic regulator in this case would 
be to ensure that the charges applied to irrigation 

customers fairly reflect the costs of service provision. 

In effect, the regulator would need comfort that 

the revenue stream from the project is sufficient to 

cover all costs, including capital costs, to avoid the 
monopolist passing costs onto other customers. It 
would also need assurance that the revenues are not 
excessive and thus represent a transfer of rents to 
the service provider. 

The Northern Adelaide Irrigation Scheme is a 
$156 million project that treats 12 Gigalitres of 
wastewater at the Bolivar wastewater treatment 
plant and distributes it to irrigation customers on 
the Northern Adelaide Plains. The business case for 
the project was based, in part, on the environmental 

benefits of reduced wastewater discharge to the 

Gulf of St Vincent and the economic and social 

benefits of intensified horticulture in the region. 

Expected water demand to support the project was 
assessed by a survey of potential users in 2015 (see, 
ARRIS 2015). 

Prices for recycled water in South Australia are 
regulated by ESCOSA, the economic regulator. 
This occurs because the provider of recycled water 
(SA Water) is a fully integrated monopoly in the 
state and holds market power. ESCOSA’s approach 

to regulating prices for recycled water is ‘light 

touch’ and limited to ensuring the nine principles 
embodied in the National Water Initiative pertaining 
to pricing recycled and stormwater are adhered to. 
SA Water publishes its pricing schedule for NAIS and 
announced a review of pricing to enhance customer 
uptake in 2022. 

1 There are complications from including employment impacts, especially if they are added to only one part of the cost-benefit analysis. 


2 Hone et al. (2022) note that the routine choice of a 7% discount rate gives rise to a variety of concerns. 


Case analysis: 

The insurance premium for 
non-rainfall dependent water 

In the midst of Australia’s millennium drought many 
towns and cities were confronted with dwindling 
water supplies. Moreover, at that time, predictions 
about the impacts of climate change pointed 
to increased variability of rainfall, including the 
prospect of more severe and extended droughts. In 
mid-2007 Melbourne’s freshwater storage fell below 
30 per cent and the spectre of a city of almost four 

million people facing significant water shortages was 

real. The Age newspaper was publishing a daily clock 
predicting the point at which water supplies would 
be exhausted. 

This provided the background to the state 
government announcing the construction of one of 
the world’s largest desalination plants. The Victorian 
desalination plant had a capital cost of $3.5 billion 
(2009 dollars) and was developed as a public-privatepartnership 
with AcquaSure. 

The announcement about the construction of the 
desalination plant was accompanied by decisions to 
enhance the connection between Melbourne’s water 
supplies to irrigation water held in catchments north 
of the city (the North-South pipeline) and to expand 
wastewater recycling. 

Prior to these announcements, water shortages were 
being managed by restrictions being placed on water 
use along with a range of incentives for water users 

to adopt more water efficient appliances. Arguably, 
these earlier approaches could be justified on 
standard cost-benefit terms, inasmuch as they had 
the benefit of forestalled major capital expenditures 

and the demand management measures had gained 
broad acceptance by the community, at least initially, 
suggesting the costs were not too onerous. 

In contrast, the desalination plant embodies distinct 
insurance characteristics, perhaps representing the 
risk aversion of decision makers at that time. 


These insurance components are most overt 
in the contracting arrangements between the 
Victoria government and AcqaSure. The payment 
to AcaquaSure has two components: (1) a water 
security payment which covers the cost of design 
and construction along with operational and 
maintenance costs, regardless of any water supplied 

(2) a water usage payment that is tied to the volume 
of water ordered by the Minister of Water on 1 April 
of each year. Water orders can take the form of 0, 50, 
75, 100, 125 and 150 Giglitres. 
Since commissioning in 2012, orders of water 
from the desalination plant began in 2016-17 
with 50 Giglitres and peaked at 125 Giglitres each 
year between 2019-20 and 2020-21. The cost to 
consumers of the water security payment where no 
water is ordered has been estimated at about $608 
million per year (Australian Cost Engineering Society, n.d.). 


Case analysis: 

Trade waste management by 
wastewater utilities 

Australian jurisdictions already have experience 
managing waste streams in sewage using a 
combination of centralised and decentralised 
institutions. This is commonly expressed in the form 
of trade waste agreements between customers 
who produce liquid wastes and the operator of the 
receiving wastewater treatment plant. The legislation 
that covers these types of agreements can vary 
across jurisdictions; for instance, in regional NSW 

liquid trade waste is defined under local government 

regulations, whereas in Victoria the Water (Trade 
Waste) Regulations sit within the state’s Water Act. 

The rationale for managing trade waste differently to 

household and ‘general business’ sewage is that the 
waste stream is of sufficient quantity and/or quality 

that it represents a risk to the management of the 
centralised treatment plant. It is also large enough 
to warrant individual metering. At a technical level, 
the local waste stream requires more treatment and, 
at an economic level, this entails additional cost 
that would otherwise be borne by the wastewater 
treatment plant. Most jurisdictions follow similar 
protocols but there can be marked variations around 
key elements, even between wastewater utilities in 
the same state. 

In order for a customer to discharge trade waste to 

the sewer network they must first secure approval 

from the plant operator (i.e. Council, Water Utility). 
A risk assessment usually follows to ascertain the 
potential impacts of the wastewater stream on 
the sewer network and wastewater treatment 
plant and this is then built into an agreement 
with the customer. 

Key parameters of interest include the volume of 
waste, the Biochemical Oxygen Demand, Suspended 
Solid, Total Kjeldahl Nitrogen and Total Phosphorous. 
For some trade waste customers, their agreement 
commits them to generating wastes that do not 

exceed specific biological and chemical thresholds. 


A charge is levied on the volume of waste that lies 
below those thresholds with additional fees applied 
when non-compliance occurs. There may also 
be charges levied to cover inspection and monitoring 
activities. Repeated breaches can result in 
forcible disconnection from the sewer network and 
prosecution. 

In simple terms, the impact of this regulatory 
approach is to shift the onus for choosing and 
managing a wastewater technology to the 
decentralised customer base. Obviously, this initially 
reduces costs for the centralised operator but it 
can be a sizeable undertaking – in SA, trade waste 
customers represent around 1.5% of all customers 
(approximately 10,000 customers). 

In order to bring compliance across such a large 

number of customers requires a significant 

monitoring and enforcement investment. In some 
cases, this is complicated by the volume of technical 
information expected to be read and understood by 
customers. SA Water, for instance, provides almost 
50 fact sheets on its website to help customers 
navigate this relationship. Accordingly, while it is 
feasible for these regulatory arrangements to offer 

some flexibility as new knowledge and technology 
emerges, significant resources are still required to 

manage any transition. 


Acknowledgements 

This work was supported by UniSA Business, 
University of South Australia, CRC-SAAFE and the 

Commonwealth Scientific and Industrial Research 

Organisation (CSIRO). 

www.csiro.au/amr