6. Conclusions
Generally speaking, the breadth of data in the LBD enables significant advances to be made in many areas of microeconomic analysis. This paper represents a first attempt by researchers to exploit some of the potential of the dataset. Given the sheer scale of the data available, we have really only scratched the surface.
The dataset stands out for both its comprehensive coverage of firms operating in the New Zealand economy (having been built primarily around government-administered data collections) and the variety of variables captured. There are distinct advantages from having "full" coverage of firm-level outcomes. For example, this paper has deliberately steered away from tables of means and, instead, focussed on complete distributions, allowing us to see how discussions of average industry labour productivity need to be nuanced by the fact that many firms in "high performance" industries actually have inferior performance to many firms in "low performance" industries (see Bartelsman et al. 2006 for an application of this line of reasoning at the cross-country level). Perhaps more importantly, future industry performance may be related to the dispersion of current firm-level outcomes (eg, if learning effects are more easily transmitted between firms that initially have similar production technologies). The coverage of the LBD makes investigation of such economic models possible.
On the variable side, key strengths of the dataset include the integration of:
- IR10 & AES data enabling the construction of a much wider set of financial performance metrics and, consequently, a more nuanced view of firm success. For example, high profitability & high productivity are not constant companions (Table 7) and decision-makers in firms may be more concerned about the former than the latter;43
- LEED variables, enabling better estimation of labour inputs and (aggregate) worker characteristics;
- Customs & IR4 data allowing superior identification of the international linkages of firms (respectively, goods exporting & foreign-ownership/foreign income); and
- A wide variety of sample surveys, expanding the set of research questions that can be tackled with the data.44
There is still work to be done with the administrative data to convince the (appropriately) sceptical reader of the validity of derived research outputs. In this regard, the caveats section of this paper has highlighted several key areas for the further development. However, we should try to avoid becoming paralysed by inconsistency in the data. The fact that we have many instances where we have multiple measures of the "same" thing (to the point where, in cases such as wages & salaries, there are four "independent" estimates within the data – AES, IR10, LEED, BOS) is a gift, not a curse, and gives us more choices in our models & our research design. The LBD gives us the advantage of seeing how those choices – choices not available to prior researchers – affect our findings.
At this stage, the analysis presented provides motivation for more work, not definitive answers to questions. This paper is a starting point for more detailed analysis of the dataset. The ongoing research programme is focussed on topics close to the heart of the public policy debate on the economy, particularly improving our understanding of the determinants of New Zealand firm performance, including the impact of government assistance to firms.
The primary weakness of the LBD is the short timespan of the data relative to studies that rely purely on BF/BAI data. Only future years of data production and integration by SNZ can correct this shortcoming. We would encourage other policymakers and researchers to consider whether they have potential uses for the LBD. A larger community of users will deepen our understanding of the New Zealand economy, and can only encourage SNZ to invest further in the development of the LBD.45
Back to Top