National Research Data Programme – NRDP

Economic Case for Data

In 2008, the consultancy Charles Beagrie Ltd published the seminal model for understanding the economic impact of well-managed research data and well-trained researchers. Developed for the Joint Information Systems Committee (JISC) in the United Kingdom (UK), the Beagrie “Keeping Research Data Safe” model has now been applied across multiple European research systems (UK, Ireland, Denmark, The Netherlands, and the EU) as well as in the USA (National Science Foundation), Canada and Australia over the last 5 years. Keeping Research Data Safe includes detailed tools for cost / benefit and value chain analysis across the life cycle of research data. In 2014, Australian economists John Houghton and Nicholas Gruen used the Keeping Research Data Safe model to show that Australian research data is worth between AUD1.9bn and AUD6.0bn per annum to the Australian economy at current levels of research expenditure.

Rather than replicate this work, this proposal takes the initial step of mapping Houghton & Gruen’s analysis to the New Zealand context, and then drawing parallels for the New Zealand research system. This means that the economic estimates we include are only indicative – without a detailed NZ study it is difficult to be exact about how much research data is curated and shared, or quantify the inefficiencies that stem from lack of data skills or lack of access to data in research. A next step would be to commission detailed analysis of the New Zealand case; however we consider that step would form an early task for the proposed NRDP.

In Australia, Houghton & Gruen identified two major sources of value from research data:

the current value of research data, generated each year through existing public research investments on a per annum basis, and

the potential value of research data management, of meta-data catalogues, standards and tools, collections, curation and sharing that could be realised from data related investments focused on sustaining a stock of research knowledge.

In developing these differing sources of value, our analysis defines an upper bound based on total direct government support for research, per annum; and a lower bound, which is defined as the researcher labour-only costs in publicly funded research, per annum. These bounds are based on figures from the National Statement of Science Investment (2015), on sector assumptions drawn from international practice and set out in Table 1, and on economic analysis conducted for us by PricewaterhouseCoopers (PwC) in Wellington.

TABLE 1.

KEY NZ RESEARCH SECTOR ASSUMPTIONS

FIGURE10.

THE VALUE OF BETTER RESEARCH DATA MANAGEMENT IN NZ UPPER BOUND FIGURES ONLY

The Current Value of Research Data

Houghton & Gruen present two alternative approaches to valuing data generated through public investment each year; the Use Value approach values data based on the time and cost of producing it, while the ROI on Research Data Activities values the likely return on researcher time.

The Use Value of Research Data
The value of anything when you first purchase it (i.e. an ice cream, a car) is assumed to be equivalent to the cost of the item at the time of purchase. The same can be said of research, and of research data – its value is equivalent to the cost of the activities involved in producing it each year. Analysis in Australia and the UK suggests that researchers spend approximately 46% of their time creating, manipulating and analysing data. Applied to the New Zealand research sector, the activity cost or Use Value of research data generated each year in New Zealand through publically funded research is 46% of total expenditure. That’s $283m pa if we take the lower bound, or $589m pa if we base our estimate on all public research expenditure (upper bound).

The ROI on Research Data
The Government invests annually in publicly funded research on the basis that the investment will return benefits to New Zealand society and economy over the long term. Return on Investment (ROI) models differ in parameters for different countries, however the underlying principles are widely understood. As the proportion of total research funding that is spent on creating, collecting, and analysing research data each year is relatively large, we can apply similar ROI calculations to understand the potential returns to New Zealand that might be generated by our annual investment in research data over the long term.

Houghton & Gruen note that in a global research community, only a portion of the value of research activity will actually accrue to the funding country, which are termed localisation returns. If we apply international norms for localisation (66%) and returns on research investment (40%) to the New Zealand context, the ROI on publicly funded research data would be between NZD235m (lower bound) and NZD490m (upper bound) per annum (NB: these figures are lower than the Use Value as we have depreciated the value of research data over only 10 years).

A more conservative approach is to produce a worst case scenario for returns on research data in New Zealand by selecting alternative values for New Zealand. To produce the most conservative estimate of the ROI on research data, we applied a New Zealand Treasury return on research investment rate (17%) which was developed in 2006 to assess R&D productivity in the NZ agricultural sector between 1927 and 2001. We also took into account expectations for the localisation of returns to research effect (55%) in a small country. Even as a worse case scenario, the ROI on 1 year’s investment in research data is at least $56m pa and possibly as high as $117m pa.

In developing our ROI analysis, we worked with PwC in Wellington; who noted that public investment in research and research data is not a one-off – we make a similar investment every year, and the returns on those investments will compound in any particular year. Again using the worst case scenario parameters, the present value of a programme of improvement for publicly funded research data is still quite significant at $110m – $230m . In calculating these returns, we assumed that improvements would happen gradually over 10 years. If we consider this investment in the optimistic terms of the international model, then compound returns over 10 years of Government investment in research data could be between $461m – $961m.

It seems clear that the proportion of annual public research funding that is devoted to collecting, creating and analysing data is considerable, and that the potential returns on this investment over the long term are substantial – even in a worst case scenario. The concern is that, as data-intensive discovery and digital research methods become the norm for global research, our New Zealand research sector find it increasingly difficult to fulfil its potential on the returns to society and productivity that we need.

The Potential Value of Data Related Investments

Houghton & Gruen suggest the potential value of research data related services and the supporting infrastructure can be estimated through assessing the efficiency gains in the research process which they create (I.e. reduced duplication, use of time, increased collaboration) and the bearing increased access to data has on the effectiveness and impact of research.

Efficiency Gains – research funding & data first use
Suppose we assume that all researchers in New Zealand have efficient access to data (or at least meta-data) produced by other NZ researchers, or data held by Government or industry, that could be made available for research. What opportunities could be found, what more could be done in such a situation? Efficient access to data or meta-data in Australia and the U.K. has been shown to offer two major advantages to research; the first is that researchers can spend significantly less time seeking (or recreating) data. The second is that research scope and scale that wouldn’t otherwise be possible, either due to a lack of resources or a lack of access to data, can now be tackled by a broader cohort of the research community. In economic terms, individual researchers and research institutions become more productive if they have efficient access to data or meta-data. Based on U.K. and Australian experience, efficient access to data can reduce time and cost of research activity by approximately 37%, suggesting a potential productivity gain in the New Zealand research sector of between NZD105m and NZD218m per annum.

Based on UK and Australian experience, efficient access to data can reduce time and cost of research activity by approximately 37%, suggesting a productivity gain in the research sector of between NZD105m and NZD218m per annum.

Benefits from Additional Data Use – active data bridges & data reuse
While this proposal suggests lifting the gains from the first use of new data will have the greatest impact on research, additional use of data has an economic value of its own. Where access to data collections and repositories, along with researcher skills development, enables researchers to use data they would otherwise be unable to obtain on their own, we assume additional research is produced that could not otherwise have been done (or a similar sized increase the scale and scope of existing research). Direct extrapolation from Australian calculations suggests the benefits of additional use in New Zealand are likely to be between NZD54m and NZD113m pa.

While these figures are extrapolations from other countries, some with more advanced data-related infrastructure and deeper R&D capital stock, simply adding the efficiency gains and benefits from additional data use suggests that system-wide services such as meta-data cataloguing, data repositories, national collections, researcher skills development and related data infrastructure have a potential value to New Zealand of between $235m and $490m per annum.

There is already evidence of good research data management in New Zealand, but we are a long way from our potential. The value of a national approach to research data is that our system could make coordinated progress. If for example we assume that we are (optimistically) 20% efficient and managing our research data on an institution by institution basis today, then a programme such as the NRDP could help lift good data management at each institution over time and deliver additional returns to research productivity of $188m – $392m per annum.

TABLE 2.

SUMMARY OF RESULTS

BACK TO TOP     |      NEXT SECTION