Research and Development Survey
Research and development (R&D) statistics include the level of research and development activity, and related employment and expenditure across all sectors of the New Zealand economy.
Purpose
The Research and Development (R&D) Survey is a joint survey between the Ministry of Business, Innovation and Employment (MBIE) and Statistics NZ. It is the main source of aggregate information on R&D expenditure in New Zealand, including information from universities, Crown Research Institutes, and government departments as well as firms.
Content
The R&D Survey collects information on current and capital expenditure and employment in R&D as well as high-level information on funding sources and the purpose(s) of the research.
Coverage
The survey is conducted every even year and excludes industries that are deemed to be non-R&D performing (eg, wholesale and retail trade). Two additional waves of the survey were also collected in 2019 and 2021 to support the development and evaluation of the R&D Tax Incentive which became available from the 2019/20 tax year. These additional collections were restricted to the business sector.
Around 2,000 to 3,000 useable responses are collected each year. As the core purpose of the survey is the production of aggregate information, the sample is focused on known R&D performers, including recipients of government funding, though the practical implementation of this focus has varied over time. R&D activity in large firms is identified through the AFUS/BRUS indicators, with a threshold of $5000, as well as a range of other sources.
Tips and Tricks
(from the Rough Guide to the LBD)
The R&D survey asks respondents to identify total R&D expenditure, then later asks them to allocate this expenditure to a range of different expense categories, identifying the shares associated with different funding sources as well as the shares associated with different purposes. In some cases, firms are given the option of whether to provide dollar values or to allocate percentage shares of the total. Statistics NZ then uses the percentage information to derive dollar values, which are identified in the dataset by the prefix “D” or “LD.” These derived variables also adjust responses to be GST-exclusive and, where respondents indicate that they are reporting for a part-year, pro-rate the responses. As with BOS, pro-rating probably doesn’t make sense for micro analysis.
Edits and imputation flags are provided for each response item, identified by the suffixes _e and _i respectively.
From 2006 onwards, response codes are used to identify unusable surveys, as discussed in section 2.10 of the Rough Guide with respect to the BOS survey. However, in the R&D survey (and other Statistics NZ surveys), F refers to firms which failed to return a response. An additional code used is K, which indicates that key questions have not been answered. Earlier years used alternative indicators (response_flag in 2002, non_response_reason_code in 2004).
Response items are also allocated an “imputation status” flag (_is), which identifies whether the observation is valid for use as a donor (V = a valid entry for use in imputing other units, U = unlinked, excluded from being used to impute other units). In the case of imputed items, the imputation status provides information on the imputation method used (M = current mean method, D = donor method, P = pro-rate method). If a survey form was identified as unusable (response code = F, Q or C), no imputation is performed.
BOS and the R&D survey both collect R&D information, but the content and coverage differs (see page 42 of the rough Guide).
Links
R&D Survey questionnaire:
SNZ description of general methodology used to produce research and development statistics:
SNZ Information release for the R&D survey 2021: