• Home
• International Trade
• New Zealand Tariffs
• Post-2005 Tariff Review
• Import Tariffs beyond 2005
• 2. Analysis of Employment Data

• New Zealand Tariffs

• Post-2005 Tariff Review

• Media Releases
• Key Decisions
• Ministerial Report
• Import Tariffs beyond 2005
• Terms of Reference
• Archive

[Released under the Official Information Act]

• 2.1 Data
  • 2.1.1 Industry Classification
  • 2.1.2 Tariffs
• 2.2 Manufacturing Employment by Industry
• 2.3 Manufacturing Employment by Region
• 2.4 Manufacturing Employment by Ethnicity
• 2.5 Manufacturing Employment by Gender
• 2.6 Manufacturing Employment by Household Type
• 2.7 Manufacturing Employment by Income Decile

We begin the review of tariffs beyond 2005 by looking at the effects of tariff changes in the past. An econometric approach would be the ideal way to do this, but is beyond the ambit of the project. Instead we use data from the four most recent censuses to establish whether there is any correlation between tariff changes and changes in employment - by industry, regional, ethnicity and gender.

2.1 Data

The tables and figures in this document have been derived from 1986, 1991, 1996 and 2001 Census data on hours worked in manufacturing industries. The data provided details on the number of workers in each manufacturing industry according to the number of hours per week that they worked. This was grouped into 1-4, 5-9, 10-14, 15-19 ... etc up to 70-74, with a final 75+ group. This data was translated into the number of hours worked by assuming the mid-point in each group represents the average number of hours worked per week for that group. Those in the final 75+ group were assumed to average 79 hours per week. Data on hours worked in each Territorial Local Authority (TLA) area in each manufacturing industry was provided. In addition, data on hours worked by ethnicity in each manufacturing industry was provided.

2.1.1 Industry Classification

For 1986 and 1991 data was provided for the NZSIC major division 3 category at the 3-digit level. 1996 and 2001 data was provided for the ANZSIC division C category, also at the 3-digit level. The correspondence between the two classifications appears in Table 2.1. Note that this correspondence is by no means exact - but is, in our judgement, robust enough for the practical purposes of this investigation.

However, note that, where at all practical, the finer ANZSIC classification was used in intermediate calculations - thus the following correspondence was only applied to provide a full 1986 to 2001 comparison.

2.1.2 Tariffs

For 1991, 1996 and 2001 tariff rates were calculated using Statistics NZ trade data on duty paid and VFD total imports by ANZSIC category.

For 1986 tariff rates were inferred from 1986/1987 input-output data, but at a more aggregated industry level. The matching of 1986 rates to the finer categories is relatively imprecise and so results should be interpreted with care. Note also that the implied 1986 data for tobacco and petroleum led to what we considered to be "rogue" changes between 1986 and 1991 and so was not used.

Table 2.1: Industry Classifications

2.2 Manufacturing Employment by Industry

Hours worked in manufacturing declined substantially between 1986 and 1991, with a fall of over 25%. Since 1991 manufacturing hours have stabilised - having risen by a marginal 1.9% between 1991 and 1996, but retreated by 3.7% over the 1996 to 2001 period.

The greatest reductions (in % terms) were recorded in industrial and other chemicals, followed by footwear (industries 351, 352, 324) - all with declines of over 80%. However, the changes for industries 351 and 352 are likely to be primarily attributable to changes in classification, given the opposite change in industry 353 and 354. The next largest changes were in tobacco, clothing, rubber, glass and paper (industries 314, 322, 355, 362, 341), all recording declines of between 50% and 80%. Making up this "top ten" were transport equipment and textiles (categories 384 and 321) with reductions of the order of 45%.

Table 2.2: Weekly Hours Worked in Manufacturing

Of these "top ten" three have experienced "double-digit" %-point reductions in tariff rates over this period - namely footwear, transport equipment and clothing ex footwear (tariff rates down by 46.5, 23.8 and 19.8 %-points, respectively).

Note: In the table below the tariff rate are calculated as tariff revenue divided by the value of imports. Hence they may not reveal where the highest rates are.

Table 2.3: Manufacturing Tariffs Estimated 1986-2001

Figure 2.1: Change in Hours Worked and Tariff by Industry 1986-2001

However, there are other large employment reductions but with only small tariff changes. The scatter of employment changes to tariff changes in Figure 2.1 shows little relationship. Despite this picture, some simple relationships were attempted - with some using the data set with the outliers eliminated - but with minimal success.

Figure 2.2: Change in Hours Worked and Tariff by Industry (ex Outliers) 1986-2001

Excluding outliers (including those where employment rose) - leaves only 18 of original 29 observations - see Figure 2.2. Clearly, however, the "pre-selection" of industries does not make this a particularly robust estimation process, even though some exclusions are entirely sensible (e.g. Chemicals and Petroleum).

For a more "robust" adjustment to the data (and to allow for the overall reduction in manufacturing), the share of each industry within manufacturing was investigated.

Table 2.4: Share of Industry in Total Manufacturing Employment (%)

As Figure 2.3 - Figure 2.6 show, there is again little sign of any robust relationships, although two of the sub-periods indicate some relationship correlation.

Figure 2.3: Change in Manufacturing Employment Share and Tariff by Industry - 1986 to 2001

Figure 2.4: Change in Manufacturing Employment Share and Tariff by Industry 1986 to 1991

Figure 2.5: Change in Manufacturing Employment Share and Tariff by Industry 1991 to 1996

Figure 2.6: Change in Manufacturing Employment Share and Tariff by Industry 1996 to 2001

2.3 Manufacturing Employment by Region

Table 2.5, Table 2.6 and Figure 2.7 indicate hours worked and changes in employment by Territorial Local Authority (TLA) area.

The greatest reduction in percentage change terms in manufacturing hours worked over the 1986 to 2001 period was recorded in the Kawerau District (down 65% - accounted for by large reduction in paper), followed by Opotiki (down 59% - largely accounted for food manufacturing and clothing).

Noticeable also - see Figure 2.7 - is that the bulk of the reductions in manufacturing employment are in the Auckland and lower North Island TLAs. In the Auckland TLAs, however, there are offsetting increases in employment in non-manufacturing industries, whereas this does not occur in the lower North Island TLAs.

Large spikes in the percentage changes in hours are largely the result of small changes in TLAs with a very low activity base. For example, Queenstown-Lakes District (manufacturing only 4% of total in 2001) and Mackenzie District (manufacturing only 2.1% of total in 2001).

To adjust for these misleading percentage changes, the change in the "share of manufacturing" in each TLA was investigated - see Table 2.7, Table 2.8 and Figure 2.8.

Over the 1986 to 2001 period, the largest percentage points reduction in the relative share of manufacturing was clearly recorded by Kawerau District - where (paper) manufacturing shrank from 65% of all hours in 1986 to just over 38% in 2001. This reduction of 27% points compares with the national mean decrease of 7.6% points (from 21.5% to 13.9%).

Other noticeable reductions in the share of manufacturing clearly occurred in the Auckland areas, as well as in some of the Wellington TLAs. On the other hand, South Island TLAs largely recorded below average changes in the manufacturing share and, indeed, in some cases recorded increases in the share of manufacturing.

The largest increases in manufacturing share over this period occurred in the Wairoa and Waitomo districts, where the leather and food manufacturing (meat processing, in particular) industries respectively contributed noticeably to increased hours worked.

Table 2.9 and Table 2.10 show the mean tariff rate for each TLA. These were calculated by using the hours worked in each TLA in each industry as weights (in a chain linked manner) and applying these weights to the industry-specific tariff.

Large changes in mean regional tariffs occurred in Waitomo, Napier and Whangarei. In each case, however, these changes are not so much because the tariffs changed, but because the industry weights changed by substantial amounts. That is, there were substantial changes in employment in meat processing (Waitomo), tobacco processing (Napier) and petroleum (Whangarei).

Table 2.5: North Island TLAs Employment Summary - Weekly Hours Worked (000s)

Table 2.6: South Island TLAs Employment Summary - Weekly Hours Worked (000s)

Figure 2.7: Change in Employment by TLA - 1986 to 2001 (%)

» Actual Size Version of This Graphic (903 x 555 pixels)

Table 2.7: North Island TLAs - Share of Manufacturing in Total Employment (%)

Table 2.8: South Island TLAs - Share of Manufacturing in Total Employment (%)

Figure 2.8: Change in Manufacturing Share of Employment by TLA - 1986 to 2001 (% points)

» Actual Size Version of This Graphic (930 x 572 pixels)

Table 2.9: Tariff by Region - South Island TLAs

Table 2.10: Tariff by Region - North Island TLAs

Comparing the change in the manufacturing employment share to the change in tariff over the whole 1986 to 2001 period shows little relationship. Interestingly, however, the 1986 to 1991 period shows some relationship, but again there are clearly many influences besides tariff changes which impacted on manufacturing.

Figure 2.9: Change in Manufacturing Employment Share and Tariff by Region 1986 to 2001

Figure 2.10: Change in Manufacturing Employment Share and Tariff by Region 1986 to 1991

Figure 2.11: Change in Manufacturing Employment Share and Tariff by Region 1991 to 1996

Figure 2.12: Change in Manufacturing Employment Share and Tariff by Region 1996 to 2001

2.4 Manufacturing Employment by Ethnicity

As noted previously, weekly hours worked in manufacturing have declined, while those in the total economy have increased over the 1991 to 2001 period. Consequently, manufacturing accounted for about 14% of total hours worked in 2001; down from 21.5% fifteen years earlier. See Table 2.11.

Table 2.11: Average Weekly Hours Worked by Ethnic Group (000s)

This 7.6% percentage point reduction in the relative size of manufacturing (as measured by average hours worked) is reflected across all ethnic groups. The nearly 26 percentage-point reduction in the Pacific People ethnic group is most noticeable. The next largest reductions are in Māori and Asian groups. In contrast, the European only group experienced the least decline at 6.5 percentage-points - slightly below the overall manufacturing average decline.

This relative shift is also illustrated in the following figures - where the reduction between 1986 and 2001 in importance of manufacturing in the Pacific People ethnic group is particularly noticeable.

The 2001 composition (by ethnic group) of the hours worked within each manufacturing industry is provided in Table 2.12, with the change in this composition since 1986 listed in Table 2.13.

In terms of the share of each ethnic group in overall manufacturing hours worked, the above table show declines between 1986 and 2001 being recorded by the European only and Māori ethnic groups - the latter albeit only a small decline. This is in contrast to the relative share of Māori in the overall economy rising, further reinforcing the picture of the decline in importance of manufacturing. On the other hand, the share of the Asian group has increased markedly across both the overall economy as well as manufacturing in particular.

Figure 2.13: Hours Worked in Manufacturing as a % of All Hours Worked by Ethnic Group - 1986

Figure 2.14: Hours Worked in Manufacturing as a % of All Hours Worked by Ethnic Group - 2001

Amongst the individual industries, the changes of particular interest relate to the noticeably large declines for the European group in food manufacturing, clothing (ex footwear), footwear, paper, other chemicals and plastics. Of note at the other end of the spectrum is the large increase for this group in transport equipment. For the Māori ethnic group however, the most noticeable movements were in clothing (ex footwear) and footwear - both down - and the increase in the leather industry. Changes worthy of note for the Pacific People group were the increase in footwear, but the contrasting decline in transport equipment. Standing out for the Asian group, are the large increases in both the clothing (excl footwear) and footwear industries - and also in electrical machinery.

Little clear pattern or relationship in the change in importance of each ethnic group and the change in tariff was present however. As shown in the charts in Figure 2.15, there is little to relate the movements across the industries to tariff changes experienced by those industries over the 1986 to 2001 period.

Table 2.12: Composition of Manufacturing Hours Worked (000s) by Ethnic Group

Table 2.13: Changes in Composition of Manufacturing Hours Worked by Ethnic Group: 1986-2001

Figure 2.15: Changes in Hours Worked by Ethnicity v Tariff Changes

Note that the above pictures, as before, exclude tobacco and alcohol.

Calculating shares across the other dimension - i.e. the importance of each industry within manufacturing by each ethnic group - does not add much. Listed in Table 2.14 below are the percentage-point changes in these shares between 1986 and 2001. Looking at the ethnic-specific changes provides little evidence that one particular ethnic group has shifted predominantly more than any other. The industry worthy of note - 322 clothing ex footwear -shows a reduction in importance being recorded across all ethnic groupings, with the noticeable exception of the Asian group. For both textiles and footwear the reductions are spread across all ethnic groups.

Table 2.14: Change in Industry Share of Manufacturing Hours Worked by Ethnic Group

2.5 Manufacturing Employment by Gender

Table 2.15: Weekly Hours Worked in Manufacturing (000s)

The gender composition of manufacturing employment has changed only marginally over the 15 year period, with female hours worked accounting for a slightly lower proportion of total manufacturing hours worked in 2001 than in 1986. (i.e. 25.6% - down from 27.1%).

This is in contrast to the shift at the total economy level where the proportion of female hours worked has risen substantially to nearly 40% in 2001, from under 35% 15 years earlier.

The detail of hours worked by gender within each manufacturing industry follows in Table 2.16.

Table 2.16: Manufacturing Hours Worked (000s) by Gender by Industry

» Actual Size Version of This Table (768 x 586 pixels)

In percentage terms the largest declines in female hours worked are recorded in petroleum (down 91.7%, but probable classification problems there); footwear (down 85%); tobacco (down 74.5%); rubber (down 71.6%); clothing excl footwear (down 64.2%); iron and basic steel (down 55.9%); with transport equipment, paper and paper products and leather also all recording greater than 50% falls.

Figure 2.16: Change in Female Hours Worked and Tariff by Industry (ex Outliers) 1986-2001

As previously found however, the large employment reductions are not clearly related to industry tariff changes. The scatter of female employment changes to tariff changes Figure 2.16 shows little relationship. Note that, as previously, large outliers attributable to spurious data (as in tobacco) and/or classification complications (as in chemicals and petroleum) are not included in Figure 2.16.

To allow for the overall decline in manufacturing hours worked, the share of female hours worked in each industry is also provided in Table 2.17.

Table 2.17: Female Share of Manufacturing Hours Worked (000s) by Industry

This shows, largest (percentage-point) declines in the female share of hours worked were recorded in leather (down nearly 20%-points); prof, scientific etc (down over 10%-points); other manufacturing (down nearly 9%-points); electrical machinery (down 6.6%-points) followed by clothing excluding footwear (down 6.1%-points).

Again though, any relationship between these changes and industry tariff changes was difficult to establish - Figure 2.17.

Figure 2.17: Change in Female Share of Manufacturing Hours Worked and Tariff by Industry (ex Outliers) 1986-2001

2.6 Manufacturing Employment by Household Type

As shown in Table 2.18 the 7.0% percentage-point reduction in the relative size of manufacturing (as measured by average hours worked) is reflected across all household types with the exception of the (extremely small) couples aged 65+ with children category. While the 9.5 percentage-point reduction in the solo parents (aged <65) with children group is most noticeable, this was also a "small" group in terms of employment hours worked. The two largest household groups (couples only, at least one aged <65; and couples, at least one aged <65, with children) on the other hand experienced percentage-point reductions fairly much in line with the overall picture for manufacturing as a whole.

However, the story for these two household type categories provides an interesting contrast. While the importance of couples (one <65) with children has declined by about 6.6%-points, this occurred as a result of an absolute fall of the order of 35% in hours worked in manufacturing by this household type - compared to a more moderate 5% fall in hours worked across the whole economy by this household type. In contrast, the importance of couples only (one <65) declined by about 7%-points as a result of hours worked in manufacturing by this group remaining fairly stable - compared to a surging 60% rise in hours worked across the whole economy by this latter group.

The similar relativities in importance of these two main groups are depicted in the following figures - where the reductions between 1986 and 2001 can be seen to be noticeably close to that of the overall average for total manufacturing. Furthermore, the decline in the solo parent household type hours worked in manufacturing compared to hours worked elsewhere is also noticeable - although, as stated earlier, this was a "small" category in terms of overall hours worked.

Table 2.18: Average Weekly Hours Worked by Household Type (000s)

Figure 2.18: Hours Worked in Manufacturing as a % of All Hours Worked by Household Type - 1986

Figure 2.19: Hours Worked in Manufacturing as a % of All Hours Worked by Household Type - 2001

The 2001 composition (by household type) of the hours worked within each manufacturing industry is provided in Table 2.19, with the change in this composition since 1986 listed in Table 2.20.

In terms of the share of each household type in overall manufacturing hours worked, the above table shows by far the largest decline between 1986 and 2001 being recorded in the couple (one <65) with children group - with the largest rise in the couple only (one <65) group.

Furthermore, there appear to be few noticeable differences amongst the individual industries - with the picture for overall manufacturing itself being a mirror of the movements recorded for the total economy. In addition, as for the analysis in other characteristics, there is little to relate the movements across the industries to tariff changes experienced by those industries over the 1986 to 2001 period - as shown for the main household types in the charts in Figure 2.20 below. Again tobacco and alcohol have been excluded.

Calculating shares across the other dimension - i.e. the importance of each industry within manufacturing by each household type - sheds little light. Listed in Table 2.21 below are the percentage-point changes in these shares between 1986 and 2001. The biggest industry mover - 311 food manufacturing - illustrates the changes as spread across all household types, with the couple only 65+ category falling the greatest. Another industry worthy of note - 322 clothing ex footwear - indicates declines across all household types with the exception of the couple 65+ with children (although caution when investigating this category given the small numbers involved).

Table 2.19: Composition of Manufacturing Hours Worked (000s) by Household Type

» Actual Size Version of This Table (843 x 598 pixels)

Table 2.20: Change in Composition of Manufacturing Hours Worked (000s) by Household Type: 1986-2001

» Actual Size Version of This Table (843 x 598 pixels)

Figure 2.20: Changes in Hours Worked by Household Type v Tariff Changes

Table 2.21: Change in Industry Share of Manufacturing Hours Worked (000s) by Household Type: 1986-2001

» Actual Size Version of This Table (843 x 598 pixels)

2.7 Manufacturing Employment by Income Decile

As noted previously, hours worked in manufacturing have declined, while that in the total economy have increased over the 1991 to 2001 period. Manufacturing accounted for 14.1% of total hours worked in 2001; down from 17.3% ten years earlier.

Table 2.22: Average Weekly Hours Worked by Family Income Decile

This 3.3% percentage point reduction in the relative size of manufacturing (as measured by average hours worked) is apparent across all 10 income decile groups. In particular however, the 5.1 percentage point reduction in the decile 2 group is most noticeable. The next largest reductions are in the decile 3 and in the decile 10 groups. In contrast, the decile 6 income group experienced least decline at only 1.9 percentage points - well below the overall manufacturing average decline.

The reduction between 1991 and 2001 in the importance of manufacturing in the 2nd and 3rd deciles is clearly evident in the following figures.

Figure 2.21: Hours Worked in Manufacturing as a % of All Hours Worked by Income Decile - 1991

Figure 2.22: Hours Worked in Manufacturing as a % of All Hours Worked by Income Decile - 2001

The 2001 composition (by income decile) of the hours worked within each manufacturing industry is provided in Table 2.23 with the change in this composition since 1991 listed in Table 2.24

In terms of the share of each decile in overall manufacturing hours worked, Table 2.24 shows the largest declines between 1991 and 2001 being recorded in the 1st and 2nd deciles - with the largest rise in the 6th family income decile.

Amongst the individual industries however, there was a wide disparity in movements, with little indication of any clear pattern or relationship in the change in importance of each decile and the change in tariff. As shown in the charts in Figure 2.23 there is little to relate the movements across the industries to tariff changes experienced by those industries over the 1991 to 2001 period. Indeed, amongst the four deciles experiencing the largest changes (i.e. those illustrated below), the only income category to record a more than negligible relationship with tariff changes was the 10th decile (high income) grouping - with the implicit suggestion that the relative weight of hours worked by this income decile varies inversely with variation in tariffs. Although at first sight this seems odd, it may simply be the mirror image of the relative declines experienced by the other deciles.

Table 2.23: Composition of Manufacturing Hours Worked (000s) by Family Income Decile

» Actual Size Version of This Table (958 x 557 pixels)

Table 2.24: Change in Composition of Manufacturing Hours Worked (000s) by Family Income Decile: 1991-2001

» Actual Size Version of This Table (958 x 557 pixels)

Figure 2.23: Changes in Manufacturing Decile Shares in Hours Worked and Tariffs : 1991-2001

Note that as before alcohol and tobacco have been excluded.

Calculating shares across the other dimension - i.e. the importance of each industry within manufacturing by each income decile - is similarly unenlightening. Listed in Table 2.24 below are the percentage-point changes in these shares between 1991 and 2001. Looking at the decile-specific changes here, provides little evidence that one particular decile has shifted predominantly more than any other. The biggest industry mover - 311 food manufacturing - illustrates that the changes are spread fairly evenly across the deciles, with the 1st decile falling the greatest and the 6th decile the least, consistent with the overall manufacturing numbers noted earlier. Another industry worthy of note - 322 clothing ex footwear - also indicates sporadic, as opposed to the systematic changes being recorded across all deciles.

Table 2.25: Change in Industry Share of Manufacturing Hours Worked (000s) by Family Income Decile: 1991-2001

» Actual Size Version of This Table (958 x 541 pixels)

⁵1986 tariff data unreliable so assumed same as 1991. Includes excise duties.

⁶1986 tariff data unreliable so assumed same as 1991. Includes excise duties.