Beyond carbon: Multi-attribute market design for ecosystem services

How to cite

Kark, Peleg; Friesen, Lana; MacKenzie, Ian; Quiggin, John. 2026. "Beyond carbon: Multi-attribute market design for ecosystem services." AEA RCT Registry. AEARCTR-0019204. https://doi.org/10.1257/rct.19204-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Peleg Kark* , Lana Friesen , Ian MacKenzie , John Quiggin

* Primary investigator

Registry ID
AEARCTR-0019204
DOI
https://doi.org/10.1257/rct.19204-1.0
First registered
Last updated
Status
In development
Trial dates
2026-08-24 – 2027-12-31
Countries
Australia ()
JEL codes
C72, C91, D24, D47, Q56, Q57, Q58
Registry record
http://www.socialscienceregistry.org/trials/19204

Abstract

This study investigates the relative efficiency and performance of alternative designs for multi-attribute markets. We compare credit stacking and bundling of environmental offsets where heterogeneous landholders simultaneously supply multiple ecosystem services. Our theoretical framework demonstrates that pure bundling violates Tinbergen’s rule and generates an inefficient pooling equilibrium. Conversely, we prove that offering a menu of bundles can restore the first-best separating equilibrium. To empirically test these theoretical results, we implement a double-auction laboratory experiment across four institutional treatments: (i) independent markets (unbundled stacking), (ii) pure bundling, (iii) an optimally weighted menu of bundles, and (iv) a sub-optimally weighted menu. Across treatments, we vary production complementarity and specialisation levels (with two values per parameter) across three repeated rounds per session. We provide actionable insights, grounded in our theoretical results and behavioural data, to inform the design and implementation of emerging national and international multi-attribute environmental markets.

Experimental design

This study employs a laboratory experiment conducted in a standard university experimental economics lab, with participants recruited from a broader student population. Participants will be part of a single group of eight participants for the duration of the experiment.

The main experimental task involves subjects making decisions about how many units of multiple fictitious goods to produce or purchase. Each unit produced generates a cost, and each unit purchased has an associated value. Subjects will engage in trade of each good via a double auction market design. Subjects’ earnings will be proportional to the profits they make from trading goods. Subjects will have different roles as either buyers or sellers, which are fixed for the duration of the experiment. Each type of buyer or seller varies in its production costs or consumption values.

We will consider four between-subject institutional treatments: (i) a stacking design where each good is traded in independent markets, (ii) a pure bundling design with a single bundle containing one unit of each good, (iii) a bundling menu design with two different bundles containing different numbers of units of each good (with optimal weights), and (iv) a bundling menu design with sub-optimal weights. Within each treatment, we systematically vary producer specialisation and joint-production complementarity parameters (with two different values per parameter). Each specialisation-complementarity parameter pair is repeated for three rounds per session.

To control for subject heterogeneity in our regression analysis, participants complete a standard demographic questionnaire following the main experimental task.

Design details

Randomization method
Participants will be recruited via email through the SONA system. They choose between a list of available sessions, and the session is randomised to a treatment before it is initialised.
Randomization unit
Individual
Clusters
16 groups with 8 individuals per group
Observations
128 individuals
Arms
32 individuals in the baseline (stacking) treatment 32 individuals in the single-package (pure bundling) treatment 32 individuals in the package menu (bundling menu) treatment with optimal weights 32 individuals in the package menu (bundling menu) treatment with sub-optimal weights
Primary outcomes
Total quantity of abatement (number of credits traded) for each type of credit, credit prices, and net benefits
Secondary outcomes
Quantity of credits traded and profits of each type of agent in the market (buyer/seller)

Cited by

No citation data available for this registration yet.

Keywords