Matching problems in biodiversity offset markets: a case study of the New South Wales biodiversity offsets scheme

Biodiversity offset credits in New South Wales are transacted within a regulatory environment defined by detailed trading rules and many different types of biodiversity credits that can lead to thin markets and high transaction costs. This paper describes a market designed to facilitate efficient and effective transactions. The market includes a search algorithm to identify who can exchange with whom, according to the regulatory constraints, and an online exchange tool to facilitate efficient price discovery and allocation of offset contracts.


Introduction
Markets are generally efficient and trusted institutions, compared with bartering or bilateral transactions, because they make it easier, safer and more financially advantageous to transact.While markets 'grow like weeds' (Roth 2002) for goods and services that can be owned by individuals, they do not evolve naturally for water filtering services, biodiversity, climate moderation and other ecosystem services.This is despite the high public value of these services and the feasibility of paying individuals to maintain or increase their supply.
Attracted by the advantages of markets, many governments have created mechanisms that facilitate transactions in ecosystem services.Biodiversity offset markets have been created, for example, to allow exchanges between developers who diminish the stock of ecosystems, and landholders able to offset them.They were first created in the United States of America with the introduction of the wetland mitigation program in the early 1970s.Offset schemes have emerged or are being tested in the UK (DEFRA 2013), France (Quétier et al 2014), Germany (Darbi and Tausch 2010), (Regnery et al 2013), Latin America and Australia as part of terrestrial and aquatic conservation programs (Kate et al 2004).
This paper describes the architecture of a market designed to transact biodiversity offsets in New South Wales.Following an overview of key characteristics of biodiversity offset transactions (section 2) we describe a reverse engineering approach in which the environmental objectives of government (in this instance) and the economic environment are taken as given (described in section 3), and the task is to identify the rules and processes that facilitate efficient transactions between self-interested agents (section 4).The final section highlights a number of insights from the New South Wales (NSW) experience that may be relevant to other jurisdictions where there is interest in creating markets for ecosystem services.

Economic characteristics of biodiversity offset transactions
Efficient markets establish conventions and codes of practice that govern behaviour, encourage truthful revelation of information, lead to efficient pricing, overcome coordination and scheduling problems, thicken markets, and facilitate participation by brokers, financial intermediaries and legal experts (see McMillan 2002).These processes increase value created from transactions, improve outcomes, and reduce transaction costs.Markets for ecosystem services are missing primarily because they are public goods.Public goods cannot be owned by individuals-and once produced their benefits are available to all.Under these circumstances, markets for ecosystem services fail, because there is no incentive for an individual to invest in their supply4 and there is no practical method of aggregating the value that all individuals' place on ecosystem services.Even if these incentive and valuation problems could be resolved, there are a range of other impediments to transactions, referred to as complexities, that can also cause markets to fail or operate inefficiently (see Helm 2015, Teytelboym 2019, Kominers and Teytelboym 2020) Three types of intervention are needed to design and create a market for biodiversity offsets.The first is to establish an economic environment that overcomes the public goods problem, defines the objective of the scheme and establishes the regulatory framework needed.The measures taken by the NSW Government in this respect are summarised in the following section.The second intervention is to design a mechanism that allocates resources efficiently within the economic environment and overcomes relevant transaction, policy, strategic and timing complexities.This task is referred to as market (or mechanism) design.The third intervention is to create incentives needed to ensure that promises to supply biodiversity offsets are honoured.This is an important problem (see Farrero andKiss 2002, Bardsley andBurfurd 2009), but is outside the scope for this paper.

The biodiversity offset scheme (BOS) in NSW
In NSW, the Biodiversity Conservation Act (2016) and associated regulations-collectively the BOS5 establishes the economic environment in which offsets are transacted.The BOS: defines the objective of the government ('to facilitate development projects that maximise commercial value at no-net-loss in the quantity and diversity of ecosystem/biodiversity services'); mandates participation (to address the public good problem); defines trading rules that prioritise, in the first instance, 'like-for-like' transactions; and establishes metrics to be used in measuring the type and quantity of offsets transacted.Key elements of the economic environment created by the BOS include a vegetation classification system summarised in figure 1, and credit trading rules (defined in regulations) that prioritise transactions to those that implement the no-net-loss objective (summarised in figure 2).The economic environment also defines: standardised site assessment protocols for development sites (Biodiversity Assessment Reports) and conservation sites (Biodiversity Stewardship Agreement); standardised contracts for landholders; a credit register; and a Biodiversity Offsets and Agreements Management System.

A market for biodiversity offsets in NSW
The economic environment created by the BOS has led to transactions in biodiversity offset credits in NSW. Figure 3 reports the volume, value and type of transactions observed 6 .These transactions have been executed through two mechanisms.Transactions can be facilitated through bilateral negotiation in which buyers/sellers of offsets (generally assisted by brokers) identify a limited number of potential counterparties and then enter one-on-one negotiation.A second transaction pathway involves a payment made to the government's Biodiversity Conservation Trust (BCT) 7 , allowing developers to proceed with their project while the BCT takes responsibility for securing offsets.The BCT price of biodiversity offsets is determined administratively 8 .
It is widely understood that neither bilateral negotiation nor administratively-determined prices lead to efficient exchanges.An alternative approach is to create a fit-for-purpose mechanism by applying the mechanism design methodology (see Hurwicz and Reiter 2006) and experimental economics methods (Plott and Smith 2008) to identify, test and refine the specific rules, processes and incentive structures 6 https://datasets.seed.nsw.gov.au/dataset/biodiversity-creditsmarket-sales-dashboard 7The BCT is a statutory body created under the Biodiversity Conservation Act 2016 which performs several functions under the Biodiversity Offsets Scheme. 8The BCT previously used a credit price calculator (the Biodiversity Offsets Payment Calculator (BOPC), based on an econometric model) to determine the payment a developer was required to pay to meet its offset obligations.The BCT transitioned to using its own administrative charge system on 17 October 2022.www.bct.nsw.gov.au/info/biodiversity-conservation-fund-chargesystemwww.environment.nsw.gov.au/-/media/OEH/Corporate-Site/Documents/Animals-and-plants/Biodiversity/biodiversityoffsets-payment-calculator-190635.pdfneeded to facilitate efficient transactions between agents who hold private information.In this market design process, the objectives and economic environment of the BOS are taken as given, and the task is to identify the rules and processes that facilitate efficient transactions9 .Well-designed mechanisms embody rules and processes that: lead to truthful revelation of private information, align incentives between the principal and agent, and are strategyproof10 .The mechanism created through this design process for the purpose of transacting biodiversity offsets is referred to in this paper as the Biodiversity Offset Exchange.It is defined by two components, as illustrated in figure 4: a search tool and a market matching tool that allows individual buyers and sellers to bargain 11 .
conducted by the NSW Audit Office (www.audit.nsw.gov.au/ourwork/reports/effectiveness-of-the-biodiversity-offsets-scheme), and the Inquiry into the Integrity of the NSW Biodiversity Offsets Scheme (www.parliament.nsw.gov.au/committees/inquiries/Pages/inquiry-details.aspx?pk=2822) conducted by the Legislative Council of the NSW Parliament, respectively. 11A third tool, not considered here, would be an online trading platform that would allow contracting to be executed effectively, with electronic settlement of transactions able to be completed.

Searching tool
'Technical matching' refers to the challenge that buyers and sellers encounter in finding a potential counterparty.In markets for private goods, technical matching is usually straight-forward where goods and services can be easily described, and buyers/sellers are highly motivated to select items that meet their needs/supply capabilities.In biodiversity offset markets the technical matching problem is significantly more complex and time-consuming because biodiversity is highly heterogeneous, nature is complex, and regulation rather than self-interest motivates participants to transact ecosystem goods and services.The latter creates incentives for market participants to minimise cost and/or game the rules rather than maximise welfare as is the prevailing incentive in other markets.
In the NSW biodiversity offset market, technical matching involves identifying offset credits that satisfy like-for-like trading rules.Seven criteria are considered in this matching process: (i) plant community type (1370 PCTIDs) 13 ; (ii) threatened ecological community (TEC) status; (iii) Tier (7 threat status tiers); (iv) offset trading group (in the form of a list of acceptable PCT matches); (v) 22 Interim Biogeographic Regions of Australia (IBRA) subregions; (vi) location (to implement the 100 km radius rule); and (vii) quantity (the number of credits required/available).These rules appear straight-forward, but collectively are extremely 12 https://app.powerbi.com/view?r=eyJrIjoiMWRmN2Y2ZTgtM2FmNC00MzZiLThlZTAtZWI3NTdkYWFmODQ0IiwidCI6Ijk2Z WY4ODIxLTJhMzktNDcxYy1iODlhLTY3YjA4MzNkZDNiOSJ9 13 PCTID is a numeric expression associate to a unique type of plant community.complex to implement (see figures 1(A) and (B), appendix).
The following examples illustrate some of the complexity inherent in the technical matching process.Example 1: offset rules do not always require an exact PCT match but describe groups of PCTs that are suitable offsets14 .PCTs can be included in different trading groups.If the loss of a PCT is proposed within a TEC, the offsetting credit is subject to stricter trading rules than non-TEC credits.Example 2: biodiversity credit obligations can be offset with higher-tier (more threatened) credits.Example 3: offsets can be procured from the same or adjacent IBRA sub-region, or from sites within a radius of 100 km (referred to as 'proximity zones').Example 4: if no technical matches can be identified, a less stringent set of rules (the variation rules) can be applied.Faced with these technical matching problems, developers and landholders in NSW typically employ brokers to reduce search and matching costs.Brokers are often accredited assessors who have specialist knowledge about biodiversity credits and market demand.
To address the technical matching problem, a search algorithm was created to facilitate systematic, low-cost searches of all credits listed on the public register.The search algorithm encodes like-for-like trading rules as a series of gateways (yes/no statements), conditional statements (and/or statements), sequential and hierarchical rules for threatened and non-TECs.It also includes geo-spatial capabilities to implement the BOS spatial proximity

Biodiversity market matching tool
Once potential sellers and buyers have been identified using the searching tool, the next required step is to address the second economic matching problem, to identify who exchanges biodiversity offset contracts with whom.A mechanism is efficient if no change in transaction pairings (counterparties), price, quantity, timing or other setting can be made to increase value created from all transactions.This objective can only be achieved if the rules and processes of bargaining cause buyers and sellers to truthfully reveal private information about their valuations of biodiversity offset credits.
There are several unusual features of the mechanism designed to transact biodiversity offset credits in NSW.The first is that the economic environment developed for offset trading creates many submarkets.Based on the existing Biodiversity Credit Market Sales Dashboard in NSW15 , the number of sub-markets across NSW could exceed 200.The number of sub-markets at any point in time will vary depending on the volume and type of offset credits listed on the public register.The second observation is that many sub-markets will have few participants as illustrated in table 1. Thin markets create scope for strategic behaviour, such as collusion, that if not addressed reduce the economic efficiency properties of the market, and reduce trust in the market institution.A third observation is that the boundaries of sub-markets intersect.This phenomenon occurs because the like-for-like rules create asymmetric transaction rights.The intersecting sub-market problem is illustrated in figure 5 (based on the example output from the search algorithm reported in table 1) where developer D1 can transact credits with landholder L3 (signified by D1 = L3), developer D4 can transact with landholders L3 and L2 (D4 = L3 = L2), but D1 cannot transact with L2 even though both D4 and D1 can both transact with L3.Participation in sub-markets intersect because some like-for-like rules are constructed as with 'or' statements-for example, offset credits can be sourced from the same IBRA sub-region or within a 100 km radius.When applied across the entire offset market, these asymmetric transaction rights create a large network of sub-markets with intersecting sets of participants.A final observation is that the structure of the offset market-the number of sub-markets and their intersections-will change as sites are listed and delisted.The structure of the market for biodiversity offset credits, will therefore need to be reconfigured on a regular basis.
The economic architecture of the transaction component of the offset market is based on the multiunit double auction (MUDA) see Gode and Sunder (1994), Plott (1994), and Barner et al (2005).This mechanism allows developers to place bids and landholders to place offers for biodiversity credits.This is known as a 'posted-price mechanism' .Developers can progressively increase their bids to secure the credits they require, and landholders can progressively reduce their offer prices to make a sale.The price formation process in this market is therefore dynamic, with participants able to adjust their bids or offers, compare prices, and select the best transaction opportunity.These processes facilitate efficient price formation-one necessary condition for efficient allocation of offset obligations and conservation effort.
When scaled to incorporate all the biodiversity credits listed on the public register, the intersecting boundary of sub-markets (illustrated in figure 5) creates an assembly problem because individual buyers/sellers simultaneously can participate in multiple credit sub-markets.The Biodiversity Exchange tool resolves this problem through the use of personalised trading screens each linked to a MUDAs according to the market structure defined by the search algorithm.Each personalised trading screen displays information about the type and quantity of offset credit required/offered by the relevant participant.The screen identifies all potential counterparties and competitors including information on the type and quantity of offsets offered/needed.Software links participants who have been matched through the search algorithm allowing them to place bids and offers and execute exchanges on all items displayed.Although the software needed to create this network of intersecting sub-markets is complex, it creates a simple transaction space in which landholders and developers focus solely on making offers and accepting bids.
An example of a personalised trading screen (used for laboratory-based trading sessions) is shown in figure 6.It lists the credit sites S801 and S802 that are offered by seller ID122, together with potential counterparties and competitors that are technical matches for each site (the lower panel of figure 6).Prices in this market are revealed as landholder ID122 submits offers in the space denoted 'price' (highlighted), while developers ID123 and ID121 do the same on their screens.Trading screens preserve anonymity of buyers and sellers to mitigate collusion.Trading screens also provide information about similar transactions, including information on 'best sell and buy offers' , 'last trade' , and price history that assist market participants to form bids and offers and execute transactions.
The Biodiversity Exchange market matching tool creates a central clearing house for biodiversity offset credit trades that could be implemented as either a continuous or call market.In the continuous format, the search algorithm would be activated following any new listing or delisting on the register of biodiversity credits.Alternatively, a call market would be organised around a schedule determined by the search algorithm.In both cases, any change in credit listings or delistings would precipitate a change in market structure and a refresh of all personalised trading screens.

The boundary of markets for ecosystem services
Under specific conditions markets are the preferred transaction mechanism because they are efficient and trusted institutions.Whilst markets for many goods and services evolve autonomously, the boundary of the market economy has been expanded to include some environmental services, such as point-source pollution control.Extending market mechanisms to manage ecosystem services is, however, at the difficulty end of the market design spectrum.A great deal of complex economic engineering and technology is required to identify specific rules, processes, incentives and metrics needed to translate the complexity and heterogeneity of nature into a mechanism that implements a stated environmental objective through transactions between self-interested buyers and sellers.
When applied to biodiversity offsets in NSW, the design process identifies a mechanism that includes a search algorithm and multi-unit, double-sided price discovery process using personalised trading screens.Further work will require the development of a system that would allow contracting to be executed effectively, with electronic settlement of transactions able to be completed.Although designed specifically for the NSW BOS, some aspects of the mechanism may be relevant to markets being considered in other jurisdictions, particularly those where ecosystem goods and services are highly heterogeneous.The benefits of developing a search algorithm, for example, are likely to increase where there are many classes of ecosystems/ecosystem services (an extensive type-space) and where trading rules are more complex.The advantages of creating a market matching tool are perhaps more universal.It enables a central clearing house to be created which thickens markets, institutionalises the preferred price discovery processes, and reduces transaction costs associated with bargaining.The benefits of a market matching tool are likely to increase where the rules and regulations established through the economic environment influence market structure (the intersecting sub-market problem in NSW).In these situations, a market matching tool, including personalised trading screens, allows the market to be routinely refreshed mitigating scope for default.Such tools can also accommodate 'smart' market functions where these are needed.A range of other complexities including: the spatial synergy problem (e.g.relevant to the creation of contiguous wildlife corridors); nonconvexity (i.e.viable ecosystems involve large, lumpy assets understood as the discontinuity problem in ecology); and metric problems (see Salzman and Ruhl 2000) may also need to be resolved in designing markets for ecosystem goods and services.Where relevant complexities cannot be resolved, markets cannot be expected to do all the allocation and pricing work.In which case limitations in economic theory, computational and technology constraints, and implementation difficulties will effectively define the boundary of markets for ecosystem goods and services.

Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).

Figure 2 .
Figure 2. Relationship between ecological credits and trading rules.

Figure 4 .
Figure 4. Broad architecture of a biodiversity offset exchange.

Figure 6 .
Figure 6.A seller's list of assets available for sale, including assets S801 and S802.

Figure B1 .
Figure B1.Search algorithm architecture: interaction between TEC and non-TEC status, and trading rules.Source: New South Wales Department of Planning, Industry and Environment

Table 1 .
Technical matches generated by the search algorithm.