A2   Purchasing Power Parity (PPP) exchange rates

For the UK sector-level PPP exchange rate estimates for 2002 were derived by updating estimates of 1997 PPPs in O’Mahony (1999) using OECD 1999 expenditure PPPs and output deflators for the UK and the US. For New Zealand a new set of sector-level PPP estimates for 1997 were prepared by Gerard Ypma at the Groningen Growth and Development Centre (GGDC). They comprise a mix of unit value ratios (UVRs) calculated as sales of products divided by quantities produced UVRs and expenditure PPPs adjusted for relative trade and transportation margins and for taxes. These GGDC PPPs were updated to 2002 on the basis of producer price changes at sector level between 1997-2002 in New Zealand and the US, with additional adjustments for electricity, gas and water, wholesale and retail based on updated 1999 OECD expenditure PPPs in order to make the New Zealand PPPs for those industries more comparable with UK PPPs.

The basic GGDC approach to such estimates is as follows: European Union countries are compared on the basis of unit values, etc., derived from Prodcom, which is Eurostat’s collective database of production censuses. All EU countries are compared bilaterally to Germany because it has the largest coverage. Germany is then compared bilaterally with the U.S., as are all other non-EU OECD countries.  At industry level the results are then multilateralised using an Elteto-Köves-Szulc (EKS) weighting procedure. Therefore, in order to incorporate New Zealand into this multi-country PPP dataset, New Zealand output prices were systematically compared with those of the U.S. for a selected benchmark year. The same exercise was also carried out for Australia in order to facilitate the sensitivity tests described in Section 4 of the main text.  The main sources used in this exercise are listed in Table A2.1

In an effort to develop criteria for deciding which type of PPP should best be used for cross-country sector-level productivity comparisons, GGDC researchers have recently analysed Supply-Use Tables for a number of countries to identify how expenditure prices and output prices are related. This analysis has then been used to develop a new dataset of industry PPPs for 45 industries and 25 countries for the year 1997. Time series are then applied to update and backdate over time from this benchmark year. Full details of this dataset are provided in van Ark and Timmer (2001) and van Ark, Stuivenwold and Inklaar (2003).

In order to derive time series of New Zealand-UK ALP comparisons, we use ‘constant PPPs’ (estimated for 2002 and then updated and backdated using sectoral price deflators for both New Zealand and the UK relative to the US). This approach is preferred for estimates of productivity growth rates as the underlying price deflators are explicitly designed to capture changes through time. A disadvantage is that the weights employed to aggregate prices up to total market economy level do not vary though time, in contrast to a ‘current PPPs’ approach where the basket of goods and services that is priced changes annually. However, a current PPPs approach also has disadvantages, for example, revisions and methodological changes in the OECD-Eurostat PPP programme have contributed to considerable instability in data series based on current PPPs (Lau and Wallis, 2005).

Appendix Table A2.17 - Sources for PPP estimates for New Zealand, Australia and US

3-digit Gross Output set for 1997

• -OECD STAN Database 2004
• -Statistics New Zealand, Input-output table 95/96, Table 2 Use
• -Statistics New Zealand, Rest of the Economy Survey 1996
• -Australian Bureau of Statistics, Australia, Input-Output table 1997
• -Australian Bureau of Statistics, Input-Output tables Product details 1996-1997
• -Australian Bureau of Statistics (2003), Mineral Production, Quantity and Value by State, 2001-02 and 2002-03
• -Groningen Growth and Development Centre, 60-Industry Database, October 2005, http://www.ggdc.net
• -OECD, Structural Statistics for Industry and Services

Agriculture

• FAOSTAT database, FAO prices and quantities for 1997

Mining

• -Australian Bureau of Statistics (2003), Mineral Production, Quantity and Value by State, 2001-02 and 2002-03
• -United Nations, 2001 Industrial Commodity Statistics Yearbook
• -Statistics New Zealand, ACPs by ANZSCC
• - Statistical Abstract of the United States 1999
• -1997 US Census of Manufactures,

Manufacturing

• -ICP PPPs and nominal values for basic headings for 1996 and 1999 from the OECD/Eurostat workgroup
• - OECD (1999), Consumption Tax Trends, 1999 edition, Paris
• - OECD and International Energy Agency (1999), Energy policies of IEA countries, 1999 review, Paris.
• -OECD STAN Database 2004
• -Trade Margins from Trade PPP calculations (see below)
• -Groningen Growth and Development Centre, 60-Industry Database, October 2005, http://www.ggdc.net
• - U.S. Census of Manufactures 1997
• - Mulligen, P.H. Van (2002), Quality Differences And Hedonic Pricing In International Comparisons, Ph.D. Thesis, University Of Groningen.

Utilities

• -United Nations, 1998 Energy Statistics Yearbook
• -FAO, Aquastat Database 2002
• -International Energy Agency’s Energy Prices & Taxes (2^nd quarter 2006)
• -ICP PPPs and nominal values for basic headings for 1996 and 1999 from the OECD/Eurostat workgroup

Trade

• -Statistics New Zealand, Annual Enterprise Survey - NZSIC-Based Financial Estimates and Sample Errors, 1995/96 Financial Year
• -Statistics New Zealand, Actual Retail Sales By Quarter By Storetype, http://www.stats.govt.nz/domino/external/PASFull/pasfull.nsf/0/73250640e09aeaca4c25671a0016ba13/$FILE/alltabls.xls
• -US Bureau of Census, 1997 Economic Census
• -Australian Bureau of Statistics, Wholesale Industry Australia, 1998-1999, 2.1 Selected Income items by industry
• -Australian Bureau of Statistics, Retail Industry Australia, 1998-1999, 2.1 Selected Income items by industry

Transport

• -World Bank Railway Database
• -ICAO, Civil Aviation Statistics of the World 1997
• -United Nations, Annual Bulletin of Statistics for Europe and North America 1999
• -OECD, Structural Statistics for Industry and Services
• - Universal Postal Union, Universal Postal Database 2004
• - OECD, Telecommunication Database 2003
• -Bolland, Weir and Vincent (2005), Development of a New Zealand National Freight Matrix
• -U.S. Department of Transportation, Bureau of Transportation Statistics (2002), National Transportation Statistics 2002, BTS02-08, Washington, DC, U.S. Government Printing Office, December 2002
• -Statistics New Zealand, Input-output table 95/96, Table 2 Use
• -Statistics New Zealand, Rest of the Economy Survey 1996
• - Statistics New Zealand, National Accounts 2004
• - Statistics New Zealand, NZ Statistical Yearbook 1998
• - United Nations ESCAP, Asia-Pacific Transport Database, Transport and Tourism Division
• - Institute of Shipping Economics and Logistics (1999), Shipping Statistics Yearbook 1999
• - Bureau of Economic Analysis, 1997 Benchmark Use Table
• - Australasian Railway Association-personal communications.
• - Australian Bureau of Statistics, Survey of Motor Vehicle Use, Australia, 2000 (9208.0)
• - Australian Bureau of Statistics, Input-Output tables Product details 1996-1997
• - Quantas Annual report 1997
• - OECD STAN Database 2005, rev. 2

Other Industries

• -ICP PPPs and nominal values for basic headings for 1996 and 1999 from the OECD/Eurostat workgroup
• - OECD (1999), Consumption Tax Trends, 1999 edition, Paris
• -Groningen Growth and Development Centre, 60-Industry Database, October 2005, http://www.ggdc.net
• - OECD STAN Database 2005, rev. 2

Statistics New Zealand, Input-output table 95/96, Table 2 Use

A3  Labour quality measurement

Our approach to estimating and comparing average labour quality in New Zealand and the UK is described in detail in Section 5.2 of the main text. To recapitulate, this measure was derived by benchmarking on graduate-level qualifications (where comparability across countries is at its strongest), and then using ratios of mean wages in non-graduate categories to mean graduate wages in each country as indicators of labour quality differences between the respective categories.

For the UK estimates of qualification shares at industry level were derived from Labour Force Surveys 1995-2004. Following advice from NZ Statistics, estimates of employment shares by qualification group at sector level in New Zealand were based on the NZ Income Survey (NZIS) which is believed to collect higher quality data on qualifications than the NZ Census. It is also an advantage for comparative purposes that the NZIS is based on an interviewer-administered questionnaire as is the Labour Force Survey in the UK. However, since the qualifications data in the NZIS are only available at a relatively high level of sectoral aggregation, more disaggregated sectoral estimates did have to be based on NZ Census data for 1996 and 2001. In addition NZIS data were only available for 1997-2004 so the estimated series was backdated to 1995 on the basis of rates of change between 1997-99.

Data on weekly pay in the UK Labour Force Survey and annual pay in the NZIS were also used to derive estimates of qualification-related wage differentials for full-time workers for aggregate manufacturing and aggregate market services in each country. The focus on full-time workers is necessary since we do not have access to hourly wage data in either country which would be conceptually preferable as an indicator of productivity.  These wage data for manufacturing and market services were then used to weight employment shares by qualification group in relevant sectors; for agriculture, mining, utilities and construction, employment shares were weighted by the wage differentials for aggregate market sectors. Table A3.2 below shows a fair degree of stability over time in the wage ratios for aggregate market sectors in each country, with the exception of the years 2003-04 in New Zealand when the survey data suggest a widening of the pay gap between graduates and non-graduates.

This approach constitutes a distinct advance on skill measures based on education inputs (eg, years of schooling) or attainments which make no effort to take account of productivity differences. However, the measure used here relies on two key assumptions (1) that relative mean pay by qualification group is reflective of productivity differences and (2) that graduate-level productivity is comparable across countries. Furthermore, as noted in Section 5.2 above, there are many concerns regarding New Zealand data on qualification levels and mean wages by qualifications group (partly due to small cell sizes in the surveys concerned). Hence, our estimates of relative labour quality need to be treated with due caution.

A more complex version of our labour quality measure would take account of inter-country differences in the age-distribution of workers in each qualification group since age is generally correlated with work experience and opportunities for on-the-job skills acquisition. Hyslop, Mare and Timmins (2003) point out that the proportion of New Zealand workers holding degree-level qualifications roughly doubled between 1986 and 2001. This means that recent increases in qualifications are concentrated in younger (less experienced) age groups which may tend to reduce the wage premia attached to degree-level qualifications. It is beyond the scope of this paper to explore New Zealand-UK differences in this respect in detail. However, it is worth noting that the UK has experienced similar rapid growth in the graduate share of employment since the 1980s which has persisted into the early 2000s (Table A3.1).

Appendix Table A3.18 -Employment in aggregate market sectors, analysed by qualifications category, 1995-2004

A: UK

B: New Zealand

Sources: UK Labour Force Survey, NZ Income Survey and NZ Census of Population and Dwellings.

Appendix Table A3.29 - Pay differentials by qualification categories in aggregate market sectors, 1995-2004 (Index numbers: Mean graduate pay=1)

A: UK

B: New Zealand

Sources: UK Labour Force Survey and NZ Income Survey.