Complex Economic and Social Systems
Te Pūnaha Matatini is using methods from complex systems analysis and organisational-level data sets to understand the role of innovation in productivity growth, and to assess the importance of knowledge, network, and supply-chain spillovers on firm behaviour.
The last decade has seen dramatic advances in our understanding of complex economic networks. Researchers at Te Pūnaha Matatini are applying new methods from complexity science to better understand New Zealand’s economic and innovation performance. New Zealand’s failure to close the gap in GDP with other advanced economies has been attributed to our small scale and distance from major markets, but the manner in which these factors influence the New Zealand economy’s ability to capture and benefit from knowledge spillovers is largely unexplored. Understanding the potentiality of spillovers from diversity will inform government policy and decision-making, and will assist in the evaluation of the effectiveness and impact of government policies.
Network analysis for the social sector
Network science has revolutionised internet search, online advertising, and fraud detection, and produced a step change in our understanding of the resilience of ecosystems and the robustness of financial systems. When applied to social relationships, network science has been used in many contexts to predict the influence of individuals on their friends, acquaintances, and family members. The analytics group at the Ministry for Social Development, iMSD, currently has time-stamped relationship data for its clients that reveals some of the structure of the social networks that influence children’s lives. There is an immediate opportunity to use this information to develop new network measures that would help predict the need for child welfare and protection services. In much the same way as Google ranks webpages based on the network structure of the world-wide web, such measures would rank vulnerable children based on risks that they are exposed to in their social networks. A partnership with academic researchers at Te Pūnaha Matatini is enabling iMSD to take advantage of this opportunity, allowing it to implement network measures in frontline decision making processes within five years.
Three-year outcomes: Development of network models that can be used to provide estimates of concern at both a the individual and community level. These can be available to CYF call centre staff, in addition to their existing databases and expertise, to help them to make real-time assessments as to whether individual cases need more in-depth investigation or intervention.
Norms, learning, and labour market behaviours upon childbearing
Despite reductions in gender disparities in the labour market in recent decades, women in New Zealand (and in most developed countries) still lag men significantly both in average earnings and pay relative to productivity. Much of the gap is related to childbearing. A woman’s decision about how much time to take off work when having a child, whether to change employers, and whether and how long to work reduced hours is based on many more factors than rewards in the labour market. For example, personal preferences and utility from time spent with a child are likely to play large roles, employer agreement may impose constraints, and the availability and cost of alternative childcare have been shown to matter. Social norms, values, and accepted behaviours may also influence decisions and outcomes. This project explores the importance of learning from observing in the workforce the employment outcomes of other women around childbearing.
In research beginning early in 2017, Isabelle Sin, with Gail Pacheco and Chao Li, is using longitudinal earnings data and birth records to document how women’s labour market participation and earnings evolve with childbearing. The proposed project will use similar data and build off this work. It will investigate econometrically how employed women’s labour market outcomes upon childbearing relate to the observable characteristics of the woman and her employer. It asks how do women make decisions about labour market participation around childbearing, such as how long to take off, whether to work part-time, and whether to leave their employer? Specifically, do women learn social norms or the value of certain decisions by observing others make them?
Three-year outcome: Identification of the reasons why women make certain labour market decisions around childbearing will inform understandings of the extent to which women make optimal choices given their preferences, contexts, and circumstances, and will enable development of new policy interventions.
Students are faced with a multitude of choices regarding their study, as they progress through higher education. Similarly, institutions of higher education are operating in an increasingly complex and competitive environment. Governments are simultaneously devolving more control over programmes and budgets to individual institutions while directly intervening in higher education systems in order to ensure greater economic efficiency, quality of outcome, student access and accountability. Tertiary institutions invest considerably in the provision of support services and targeted programmes with the goals of improving student outcomes (both academically and more broadly) and increasing the participation and retention of underrepresented groups. This project will quantify the downstream benefits of providing this support and will identify differences in the level of impact across different sub-groups within the student population. Large amounts of data are collected about students during the course of their education, e.g. high school and university academic records, demographic data, and records of students’ interactions with services such as university accommodation and targeted support or extension programmes. Analysing data such as these can help to improve higher education practice by improving student experiences and informing the design of programmes to increase participation and retention of underrepresented groups in higher education.
This project consists of two parallel components. The first will use linked panel data to econometrically investigate a causal relationship between student service engagement and student outcomes. The second component will use a mixed methods approach, to investigate how students’ backgrounds and science identity influence their participation and outcomes in different fields of science during tertiary education. Each of these research components is linked to a PhD project.
Three-year outcome: Identification of the impacts on student outcomes, from changes in public policy; provision of a depth of insight into and greater understanding of priority groups of students; proposal of mechanisms and approaches that can be used to boost science capital for students from underrepresented groups and thus increase their participation, retention, and outcomes in science education.
Economic Geography: The impacts of spatial and network proximity
A key insight from economic geography is that economic density enhances economic performance and resilience. Spatial proximity facilitates productive/ beneficial interactions between people and between firms. These interactions improve economic performance and social outcomes. They also contribute dynamic benefits, enhancing learning and innovation, and the ability to adapt to a changing world (resilience). Recent studies have examined non-spatial forms of network proximity that may yield benefits for firms, cities and regions that are similar to those associated with spatial proximity. The potential advantages of non-spatial proximity is of particular importance for New Zealand and its regions, where spatial proximity is low due to remoteness and low density.
Our research will focus on two aspects of economic geography – the contribution to firm and city performance of worker mobility across urban job networks, and the role of inventor and firm proximity to innovation and knowledge flows, as captured by patent citations.
Three-year outcome: Develop measures of spatial and network proximity for New Zealand firms and people, and identify the impact that different forms of proximity have on knowledge flows and economic outcomes.
The science of funding
There is insufficient evidence on the effectiveness of public research funding mechanisms. Organisations award funds to those they judge most likely to succeed, which introduces selection bias into any evaluation of subsequent success. There is also little hard evidence about the effectiveness of these selection mechanisms, which themselves consume considerable time and resources. Finally, governments are increasingly focused on trying to maximise the impact of the research that they fund, but there is little systematic measurement of the extent of those impacts. Evaluation of the impacts of public research requires data that identify funded and unfunded research proposals and link them to institutions, individuals, funding amounts, and research areas. Such datasets are only just beginning to appear around the world. In New Zealand, the RSNZ has maintained records of all of its applicants back to 2000, and of funding decisions since 2003. MBIE is now leading a new cross-ministry effort to develop a National Research Information System (NRIS), a comprehensive and detailed dataset of a kind matched almost nowhere else in the world. We have developed a framework for tabulating impacts across multiple objectives using direct measures, indicators of intermediate outcomes, or proxies as appropriate. We have also evaluated the second-round proposals considered by the Marsden Fund, the impact of public funding of private R&D on firms’ innovation, and developed network measures for collaboration across research institutions. We anticipate characterising the outputs and outcomes of research projects using bibliometric measures, as well as additional indicators collected from public sources, such as references to research outputs made in published decisions under the Resource Management Act and Environment Court.
Three-year outcome: develop econometric models of the research process, including researchers, projects, and programmes, which enable funding agencies to design policy and programme structures with reliable selection mechanisms, and which are supportive of career development, diversity of researchers, and foster collaboration.
Risk triage for regime shifts: Managing climate change and other factors
Climate change is recognised as one of the most important problems facing society and the environment. Yet, there are surprisingly few examples of successfully incorporating climate impacts and adaptation in decision-making, especially in New Zealand. One step forward for better decision-making is to better understand where and when climate change and other global environmental change factors could push natural or agro-ecosystems, and related socio-economic systems into ‘regime shifts’. Regime shifts or ‘tipping points’ are important when they imply large, sudden and potentially irreversible changes in a system. Providing analytical insights about regime shifts is a core goal of complexity and network research, and unites all three themes of TPM, by (1) utilising diverse biological, environmental and economic datasets, (2) characterising the uncertain risks of catastrophic regime shifts in nonlinear systems with (3) a focus on informing practical policy and economic decisions. An important focus of our work will be identifying the potential for co-occurring tipping points in time and space, cascading across issues or geographies previously managed as unrelated to one another. Examples will be explored that link spatial or temporal regime shifts in climate and biological data to those in economic data or representations of parliamentary debate.
Three-year outcome: development of a process for informing policy and management decision-making processes about risks of regime shifts in ecological, environmental, governance, and socio-economic systems.
Policy network theory attempts to describe the emergence of policy from the interactions between actors in the policy-making system. Twenty years ago, Dowding observed that the theory had failed to rise above metaphor. He laid down a challenge “to produce a network theory; where the properties of the network rather than the properties of its members drives explanation, political science must utilize the sociological network tradition, borrowing and modifying its algebraic tradition.” Yet he was pessimistic about this possibility: “The quality of the data is necessarily too poor for determinate predictions because collecting such high quality data requires us to know the answers to the questions we are posing.” Since Dowding made his remarks policy network theory has grown in sophistication (drawing on complexity science for instance) but descriptions of policy network theory is still substantively grounded in metaphor. Policy-makers today leave a digital record of a portion of their interactions and outputs, relieving researchers of at least some of the burden of data collection. The goal of our project is to utilise this, using modern mathematical tools to develop and test formal models of policy-making systems. In particular, we will draw on digitised records of parliamentary debate from New Zealand, Canada, and the United Kingdom; digitised records of legislation as it passes through select committee, parliament, and finally into law; and on-line media reports.
Three-year outcome: Development of a framework to evaluate policy-making, providing insight into political strategies, advocacy coalitions, and the influence of individual politicians.