Consumers price index review: 2017 (revised)

Transcription

1 Consumers price index review: 2017 (revised)

2 Crown copyright See Copyright and terms of use for our copyright, attribution, and liability statements. Citation Stats NZ (2018). Consumers price index review: 2017 (revised). Retrieved from ISBN First published in January 2018, revised in April 2018 by Stats NZ Tatauranga Aotearoa Wellington, New Zealand Contact Stats NZ Information Centre: Phone toll-free Phone international

3 Contents Revision... 6 Purpose and summary... 6 Purpose... 6 Summary of key points... 6 About the CPI... 6 About the CPI review... 7 Changes to the CPI basket... 8 Criteria for selecting basket items Updating item specifications New price collection methods Information sources for the basket review Analysing the expenditure weights Main changes to CPI weights Food Alcoholic beverages and tobacco Clothing and footwear Housing and household utilities Household contents and services Health Transport Communication Recreation and culture Education Miscellaneous goods and services CPI all groups plus interest Tradables/non-tradables CPI pricing centres and regional expenditure weights Regional pricing centres Regional expenditure weights versus population weights Calculating regional expenditure weights Reweighting the basket Background to reweighting

4 Excluding out-of-scope expenditure Price updating expenditure Volume adjustments Managing cyclical or volatile expenditure Household expenditure levels Methods for deriving expenditure weights Purchase of new housing Purchase of new motor cars Domestic air transport International air transport Alcoholic beverages Cigarettes and tobacco Insurance services Health services Education Clothing and footwear Changing the CPI base period Upcoming CPI work Household living-costs price indexes review Retrospective superlative index CPI rolling review CPI review

5 List of tables and figures List of tables 1 Items added to the CPI basket in Items removed from the CPI basket in Changes to CPI basket item specifications in Tradable and non-tradable items added to the CPI basket Information sources for out-of-scope adjustment factors Volume adjustments for CPI items List of figures 1 Consumers price index weights, group level Food, subgroup-level expenditure weights Alcoholic beverages and tobacco, class-level expenditure weights Clothing and footwear subgroup-level expenditure weights Housing and household utilities, subgroup-level expenditure weights Household contents and services, subgroup-level expenditure weights Health, subgroup-level expenditure weights Transport, subgroup-level expenditure weights Communication, subgroup-level expenditure weights Recreation and culture, subgroup-level expenditure weights Education, subgroup-level expenditure weights Miscellaneous goods and services, subgroup-level expenditure weights CPI regional expenditure and population proportions

6 Revision We revised this paper following corrections made to Household expenditure statistics: Year ended June Purpose and summary Purpose Consumers price index review: 2017 outlines the changes we made as a result of a review of the consumers price index (CPI). Summary of key points We added 15 items to the CPI basket and removed 23. Housing and household utilities, and miscellaneous goods and services had the biggest increases in weight. The weights for food; alcoholic beverages and tobacco; clothing and footwear; recreation and culture; and education also increased. Increases were influenced by higher spending on restaurant meals and ready-to-eat-food, cigarettes and tobacco, home ownership (including additions and alterations), and insurance. Transport had the biggest decrease in weight. The weights for household contents and services, health, and communication also decreased. Decreases were influenced by lower spending on petrol and telecommunication services. We reviewed our method to calculate the weight for out-patient services which caused the decrease in health. About the CPI The CPI measures the changes in prices that households pay for goods and services. Price change is measured by tracking the prices of individual items that make up a representative basket of goods and services. Each quarter we collect about 100,000 prices. We visit retail outlets such as supermarkets, department stores, and clothing shops. We send about 2,500 surveys to a range of businesses each quarter, including construction firms, medical centres, and used car yards. We visit websites to collect prices for things like streaming services, software, and private accommodation rented from others. We also collect data from various other sources, including scanner data for consumer electronics such as televisions, laptops, and mobile phones. The CPI has a broad range of users. The Reserve Bank may adjust the official cash rate based on the CPI, which may mean banks adjust their interest rates. This will affect mortgage interest rates, and returns on investments for those saving money. Other government agencies may use the CPI to adjust payments such as Jobseeker Support and Superannuation. The CPI also helps to inform wage negotiations between employers, employees, and trade unions by reflecting changes in the cost of living. 6

7 CPI infographic has more information on the CPI. About the CPI review We review the CPI every three years to ensure the index remains relevant. Household spending patterns change over time as tastes, lifestyles, and incomes change, and as the range of goods and services available for households to acquire changes. Spending patterns also change as a result of relative price change. Households tend to buy more of goods and services that become relatively cheaper and less of goods and services that become relatively more expensive. For example, if apple prices increased a lot, but pear prices increased only a little, consumers might be expected to purchase more pears and fewer apples than before. The 2017 review has reselected the basket, and updated the relative importance of the items within it. This is to reflect spending patterns for the year to June 2016 (2015/16). Previous weights were based on spending patterns for the year to June 2013 (2012/13). Figure 1 compares the new September 2017 weights for the 11 CPI groups with the June 2014 and June 2011 quarter weights. Figure 1 1 Consumers price index weights, group level Consumers price index weights Group level Food Alcoholic beverages and tobacco Clothing and footwear Housing and household utilities Household contents and services Health Transport Communication Recreation and culture Education Miscellaneous goods and services Percent Source: Stats NZ 7

8 Changes to the CPI basket As part of the 2017 basket review we have: added 15 items removed 23 items changed specifications for items. There are now 701 items in the CPI basket, down from 709 after accounting for additions, removals, and changes to product specifications. Table 1 lists the items added to the basket. Table 1 1 Items added to the CPI basket in 2017 Items added to the CPI basket in 2017 Item Fresh herbs Olives Flavoured tea Hem of trousers Food preparations Storage costs Hearing aids Physiotherapist fees Bicycle helmets Cellphone cases Headsets/headphones Pet insecticides Admission charges to zoos Private accommodation rented from others Body massages Details Have grown in relative importance. Have grown in relative importance. Flavoured and herbal tea has grown in relative importance. Repairs and alterations to clothing has grown in relative importance. Mixers, blenders, juice extractors, food processors have grown in relative importance. Have grown in relative importance. Added to improve coverage of health equipment. Have also grown in relative importance. Significant household expenditure. Added to improve coverage of medical services. Added to represent safety devices for transport. Have grown in relative importance. Added to represent cellphone accessories. Significant household expenditure. Significant household expenditure. Added to improve coverage of audio-visual equipment. Significant household expenditure. Added to improve coverage of pets and pet related supplies. Significant expenditure. Added to improve coverage of cultural services. Has grown in relative importance. Added to improve coverage of accommodation services which has grown in relative weight. Has grown in relative importance. Added to improve coverage of personal care which has grown in relative weight. Source: Stats NZ 8

9 Table 2 lists the items removed from the basket. Table 2 2 Items removed from the CPI basket in 2017 Items removed from the CPI basket in 2017 Item Alfalfa sprouts Spring onions Taro Canned corn Luncheon meat Cottage cheese Takeaway milkshakes Liqueurs (on-licence) Interior wallboard Wall paper roll Sewing machines Antacids In-car satellite navigation units Home line connection charges DVD, Blu-ray players, and player/recorders MP3 players Set-top boxes for television sets External computer hard drives Soft toys Hire of DVD discs Domestic package holidays Details Relatively low expenditure; celery and lettuce remain in the basket. Relatively low expenditure; onions remain in the basket. Relatively low expenditure; kumara and potatoes remain in the basket. Relatively low expenditure; canned tomatoes remain in the basket. Relatively low expenditure; bacon, salami, sausages, and ham remain in the basket. Relatively low expenditure; cheddar cheese, camembert cheese, and processed cheese slices remain in the basket. Relatively low expenditure; ready-to-eat, takeaway, and fast food already sufficiently represented in the basket. Relatively low expenditure; spirits, liqueurs, spirit-based drinks (all off-licence), spirits, and spirit-based drinks (both on licence) remain in the basket. Relatively low expenditure; sufficient items of property maintenance materials remain in the basket. Declining expenditure; relatively low expenditure; sufficient items of property maintenance materials remain in the basket. Declining expenditure; relatively low expenditure. Relatively low expenditure; sufficient other pharmaceutical products remain in the basket. replaced by satellite navigation applications on cellphones. Relatively low expenditure. Declining expenditure. Declining expenditure. Declining expenditure. Declining expenditure. Relatively low expenditure; dolls remain in the basket. Declining expenditure. Relatively low expenditure. 9

10 Items removed from the CPI basket in 2017 Item Prams, pushchairs Cheque-book fees Details Relatively low expenditure; declining expenditure. Abolished. Source: Stats NZ Table 3 summarises the changes to product specifications. Table 3 3 Changes to CPI basket item specifications in 2017 Changes to CPI basket item specifications in 2017 Item Chicken (whole, frozen) Beer (off licence) hosiery New vehicles Details We previously priced sizes 16 24, we now price 1.5kg or higher. We expanded the sample to include craft beers. quarterly basis from department stores. We reviewed and updated the models of cars we collect prices for: five models changed to similar but more popular models of the same make two models removed five models added, including one electric car model. Taxi fares Stamps and postagepaid packaging Cellphone services We expanded the sample to include ride-sharing services. We reviewed the sample of postage goods and services we collect prices for. We removed postage-paid envelopes (C5, bubble, fast post) from our collection as this service is no longer provided by NZ Post. We reviewed and updated the cellphone services item. Previously we collected prices for prepaid cellphone services and cellphone pricing plans as part of the same sample for cellphone services. We split cellphone services into two separate items with separate relative weights: prepaid cellphone cards and other top-ups cellphone pricing plans. International flights Kittens We expanded the sample to include five additional destinations. One to each of Australia, Pacific Islands, and Asia; and two to North America. We removed kittens as a separate item from our price collection for pets, as kitten prices are collected as part of the cats sample. 10

11 Changes to CPI basket item specifications in 2017 Item Subscriptions and donations to sports clubs International package holidays Instruction books Details We reviewed and updated the sample of sports clubs subscriptions we collect prices for: subscriptions to bowls clubs was removed subscriptions to football clubs was added The transport module in the HES has changed resulting in the removal of international package holidays and the introduction of a combined international flights and accommodation subgroup. We reviewed and updated our presentation of the international package holidays, international air transport, and international accommodation services subgroups and classes. We removed international package holidays from the CPI basket of goods and services. We added overseas accommodation prepaid in New Zealand to the basket. As a result, the relative weight of the accommodation services subgroup and the air transport subgroup increased, while the former package holidays subgroup was removed from the relative weights. We reviewed the types of instruction books we collect prices for as part of our sample. We removed two types of instruction books as we feel the item is sufficiently represented by prices collected for: health books technical books. Source: Stats NZ 100 years of CPI basket change provides an interactive visualisation of selected basket additions and removals in the CPI over the 100 years from 1914 to See Consumers price index review: 2017 for a full list of the items in the 2017 CPI basket. Criteria for selecting basket items Basket items are selected to represent the much wider range of goods and services households purchase. We include particular items in the basket to ensure there is a good representation across the groups, subgroups, classes, and sections. We select more items for classes and sections where there is a relatively high variation in price change (ie where the prices of items in the class or section tend to move differently), than for classes and sections with little variation (ie where prices move similarly). 11

12 We typically add items that have grown in relative importance to the point where they make up a significant share of household spending. We remove items that have declined in relative importance to the point where they no longer have a significant share of household spending. Another aim of the basket review is to add new goods and services that have become available since the previous review, provided household spending has reached a significant level. We may also add items that have historically made up a significant share of household spending, but have been excluded from the basket due to difficulty collecting prices and/or adjusting the collected prices to account for changes in the quality of the products (quality adjustment). Updating item specifications We made some changes to the specifications (eg models, varieties) of some items as a result of reviewing the basket (see table 3 above). However, we did not specifications as part of the 2017 CPI review. New price collection methods Between the 2014 and 2017 CPI reviews we trialled and implemented new price collection techniques from alternate data sources. By adopting collection from secondary data sources we are able to increase the scope and accuracy of our pricing while reducing respondent burden. For purchase of second-hand cars in the transport group, we now use administrative data from New Zealand Transport Agency to collect more prices and more specifications for hedonic price modelling. The September 2017 quarter was also the first time we collected prices for certain items using web-scraping and API querying. This reduces manual processing time and the chance of human error. Web-scraping also allows us to collect prices for those harder-to-get service providers such as online accommodation and transport, and private accommodation rented from others. Work is continuing on exploring new approaches and data sources for the CPI. For more information on our broader work on transforming our data collection methods, see our paper Towards a big data CPI for New Zealand, which was presented at the Ottawa Group meeting last year. Information sources for the basket review The 2017 basket review was informed by the lower-level 2012/13 and 2015/16 Household Economic Survey (HES) expenditure data, feedback from CPI price collectors, supermarket scan data from the Nielsen Company, retail transaction data obtained from market research company GfK, and information provided directly by businesses and government organisations. The HES was filled out by a sample of 3,500 households. The number of households represented by the HES sample increased about 3.0 percent, from about 1,640,800 to about 1,690,400 households. 12

13 Analysing the expenditure weights This chapter analyses changes to the CPI weights. The following analysis provides graphs of the new CPI weights, compared with 2014 and Note that the 2017 weights are for September 2017 while the 2014 and 2011 weights are for June 2014 and June Tables 1 and 2 in Consumers price index review: 2017 tables CPI expenditure weights at the September 2017 quarter. The tables also include the expenditure weights at the June 2011 and June 2014 quarters. We publish the expenditure weights and indexes at group, subgroup, and class level. These weights are fixed, down to the class level, until the next reweight scheduled for We will monitor the relative weights below the class level of the New Zealand household expenditure classification and may adjust them where necessary to reflect volume-related shifts in relative importance. Expenditure for the year to June 2016 (the weight reference period) was price updated to the September 2017 quarter (the price reference period). The effect of price updating is to calculate how much households would have to spend in the September 2017 quarter to purchase the same quantity (volume) of goods and services purchased in the year to June Expenditure on the goods and services purchased in 2015/16 is 2.75 percent higher after price updating to the September 2017 quarter. Main changes to CPI weights The 2017 CPI weights show increases (from 2014 to 2017) in the relative importance of: restaurant meals and ready-to-eat food (from 4.35 percent in 2014 to 4.93 percent in 2017) cigarettes and tobacco (from 2.30 percent to 2.63 percent) purchase of new housing (including additions and alterations (from 4.20 percent to 5.50 percent) property rates and related services (from 3.18 percent to 3.54 percent) purchase of vehicles (from 3.48 percent to 4.47 percent) insurance (from 2.50 percent to 3.00 percent). The 2017 CPI weights show decreases (from 2014 to 2017) in the relative importance of: grocery food (from 7.04 percent in 2014 to 6.60 percent in 2017) property maintenance (from 3.09 percent to 2.14 percent) household energy (from 4.54 percent to 4.14 percent) furniture, furnishings, and floor coverings (from 1.47 percent to 1.11 percent) petrol (from 5.03 percent to 4.06 percent) telecommunication services (from 3.16 percent to 2.48 percent) audio-visual and computing equipment (from 1.16 percent to 0.98 percent). 13

14 Food The food group has a September 2017 quarter expenditure weight of percent, compared with percent in When we express the food purchased in 2012/13 and 2015/16 in June 2014 and September 2017 quarter prices, respectively, spending rose percent. This increase was partly influenced by a 3.39 percent increase in food prices from the June 2014 quarter to the June 2017 quarter. The main information source we used to reweight the food group was the 2015/16 HES. However, because the HES does not provide accurate information for some food items (eg confectionery and soft drinks), we also sourced information from food manufacturers and distributors, and from supermarket scan data (from the Nielsen Company). The relative importance of the restaurant meals and ready-to-eat food subgroup increased from 4.35 percent to 4.93 percent, due to a percent increase in spending. The relative importance of grocery food decreased from 7.04 percent to 6.60 percent, due to a relatively small expenditure increase of 2.95 percent. Figure 2 compares the September 2017 expenditure weights for the food subgroups with those for 2011 and Figure 2 2 Food, subgroup-level expenditure weights Food Subgroup level expenditure weights Fruit and vegetables Meat, poultry, and fish Grocery food Non-alcoholic beverages Restaurant meals and ready-to-eat food Source: Stats NZ Percent 14

15 Alcoholic beverages and tobacco The alcoholic beverages and tobacco group has a September 2017 quarter expenditure weight of 7.11 percent, compared with 6.96 percent in The relative weight of alcoholic beverages decreased from 4.66 percent in 2014 to 4.48 percent in This is due to decreases in the relative weight for all the alcoholic beverage classes: beer and spirits and liqueurs. These decreases are due to an increase in the relative proportion of off-licence purchases to on-licence purchases reported in the HES. The relative weight of cigarettes and tobacco increased from 2.30 in 2014 to 2.63 in Prices have risen about percent since the June 2014 quarter due to regular increases in excise taxes. Increases in price were partly offset by a decrease in the volume of cigarettes and tobacco purchased. Figure 3 compares the September 2017 expenditure weights for the alcoholic beverages and tobacco classes with those for 2011 and Figure 3 3 Alcoholic beverages and tobacco, class-level expenditure weights Alcoholic beverages and tobacco Class level expenditure weights Beer Wine 2011 Spirits and liqueurs Cigarettes and tobacco Percent Source: Stats NZ Clothing and footwear The September 2017 quarter expenditure weight for clothing and footwear is 4.36 percent, up from 4.21 percent in the June 2014 quarter. Spending on clothing and footwear increased percent from the 2012/13 weights expressed in June 2014 quarter prices to the 2015/16 weights 15

16 expressed in September 2017 prices. Prices increased 1.22 percent between the June 2014 quarter and the September 2017 quarters. Figure 4 compares the September 2017 expenditure weights for the clothing and footwear subgroups with those for 2011 and Figure 4 4 Clothing and footwear subgroup-level expenditure weights Clothing and footwear Subgroup level expenditure weights Clothing Footwear Source: Stats NZ Percent Housing and household utilities Housing and household utilities remains the largest weighted group in the CPI. Its weight increased from percent in the June 2014 quarter to percent in the September 2017 quarter. The group s increase was influenced by increases in the relative weights for home ownership, and property rates and related services. These increases were offset by decreases in the relative weights for housing rentals, property maintenance, and household energy. Spending on home ownership (which includes both construction of new houses, and alterations and additions to existing houses) increased percent, which was influenced by a percent increase in prices. Spending on property rates and related services increased percent, which was influenced by a percent price increase. Housing rentals expenditure increased 9.43 percent, but due to other areas with stronger growth, the relative weight for housing rentals showed no real change between the June 2014 quarter and the September 2017 quarter (9.22 percent versus 9.20 percent respectively). 16

17 Figure 5 compares the September 2017 expenditure weights for the housing and household utilities subgroups with those for 2011 and Figure 5 5 Housing and household utilities, subgroup-level expenditure weights Housing and household utilities Subgroup level expenditure weights Rentals for housing Home ownership Property maintenance Property rates and related services Household energy Percent Source: Stats NZ Household contents and services The household contents and services group has a September 2017 quarter expenditure weight of 4.38 percent, compared with 4.65 percent in This decrease is largely caused by a percent decrease in spending on furniture, furnishings, and floor coverings. Spending on this subgroup increased 12.7 percent between 2011 and 2014, influenced by a rebound from low spending on furniture during the six-quarter economic recession that began in March Figure 6 compares the September 2017 expenditure weights for the household contents and services subgroups with those for 2011 and

18 Figure 6 6 Household contents and services, subgroup-level expenditure weights Household contents and services Subgroup levelexpenditure weights Furniture, furnishings, and floor coverings Household textiles Household appliances Glassware, tableware, and household utensils Tools and equipment for house and garden Other household supplies and services Source: Stats NZ Percent Health The health group has a September 2017 quarter expenditure weight of 3.78 percent, down from 3.94 percent in The decrease is due to a change in the method we used to estimate expenditure for some outpatient services. The method involved using HES expenditure data and supplementing it with insurance claims. When reviewing this method and comparing results with other sources, we determined the method we used before included income from non-household sources, and therefore over-estimated expenditure. Figure 7 compares the September 2017 expenditure weights for the health subgroups with those for 2011 and

19 Figure 7 7 Health, subgroup-level expenditure weights Health Subgroup level expenditure weights Medical products, appliances, and equipment Out-patient services Hospital services Source: Stats NZ Percent Transport The transport group has a September 2017 quarter expenditure weight of percent, down from percent in the June 2014 quarter. This decrease is mainly driven by petrol. The weight for petrol decreased from 5.03 percent to 4.06 percent. Total spending on petrol decreased percent, which was influenced by a decrease of percent in prices between June 2014 and September The weight for purchases of vehicles increased from 3.48 percent in 2014 to 4.47 percent in The expenditure on purchase of vehicles increased percent between June 2014 quarter and September 2017 quarter. The weight for international air transport increased from 1.45 percent to 1.82 percent, while domestic air transport decreased from 0.68 percent to 0.56 percent. Figure 8 compares the September 2017 expenditure weights for the transport subgroups with those for 2011 and

20 Figure 8 8 Transport, subgroup-level expenditure weights Transport Subgroup expenditure weights Purchase of vehicles Private transport supplies and services Passenger transport services Source: Stats NZ Percent Communication The communication group has a September 2017 quarter expenditure weight of 3.20 percent, down from 3.61 percent at the June 2014 quarter. Within this group the weight for telecommunication equipment increased from 0.29 percent in 2014 to 0.61 percent. The weight for telecommunication services decreased from 3.16 percent to 2.48 percent. Figure 9 compares the September 2017 expenditure weights for the communications subgroup with those for 2011 and

21 Figure 9 9 Communication, subgroup-level expenditure weights Communication Subgroup level expenditure weights Postal services Telecommunication equipment Telecommunication services Source: Stats NZ Percent Recreation and culture The recreation and culture group has a September 2017 quarter expenditure weight of 9.40 percent, the same as that for the June 2014 quarter. The weight for audio-visual and computing equipment fell from 1.16 percent to 0.98 percent, influenced by prices decreasing percent between June 2014 and September The weight for other recreational equipment and supplies increased from 2.23 percent to 2.40 percent, driven by a percent increase in expenditure on pets and pet-related products. We made a change to the structure of the accommodation services, and package holidays subgroups. This was a result of changes to the HES module for package holidays. The HES asked households for their expenditure on flights and accommodation instead of expenditure on package holidays. During the 2017 CPI review, we changed our approach to this by dropping package holidays from the basket and increasing the relative weights for accommodation services and air transport (however, we will still continue to price the main components of a package holiday). We added overseas accommodation prepaid in New Zealand, which took most of the relative weight previously assigned to international package holidays. People still spend for package holidays, but many are increasingly booking their own holidays and flights online. As a result, the relative weight for the package holidays subgroup dropped from 1.13 percent in June 2014 to zero in September Consequently, the relative weight for accommodation services increased from 0.48 percent in the June 2014 quarter to 1.71 percent in the September 2017 quarter. Together, the combined relative weight of accommodation services and package holidays increased from 1.61 percent in the June 2014 quarter to 1.71 percent in September The 21

22 relative weight of international air transport (which is in the transport group) increased from 1.45 percent in the June 2014 quarter to 1.82 percent in the September 2017 quarter. Figure 10 compares the September 2017 expenditure weights for the recreation and culture subgroups with those for 2011 and Figure Recreation and culture, subgroup-level expenditure weights Recreation and culture Subgroup level expenditure weights Audio-visual and computing equipment Major recreational and cultural equipment Other recreational equipment and supplies Recreational and cultural services Newspapers, books, and stationery Accommodation services Package holidays Source: Stats NZ Percent Education The education group has a September 2017 quarter expenditure weight of 2.01 percent, up from 1.91 percent in the June 2014 quarter. Much of the increase came from tertiary and other post-school education, which had increased spending of percent. Figure 11 shows the September 2017 expenditure weights for the education subgroups with those for 2011 and

23 Figure Education, subgroup-level expenditure weights Education Subgroup level expenditure weights Early childhood education Primary, intermediate, and secondary education Tertiary and other postschool education Other educational fees Source: Stats NZ Percent Miscellaneous goods and services The miscellaneous goods and services group has a September 2017 quarter expenditure weight of 8.02 percent, up from 7.28 percent in the June 2014 quarter. Most of this increase came from insurance, which increased from 2.50 percent to 3.00 percent influenced by higher prices for dwelling and health insurance. Within insurance, dwelling insurance increased from 0.41 percent to 0.60 percent, while life insurance increased from 1.02 percent to 1.20 percent. Figure 12 shows the September 2017 expenditure weights for the miscellaneous goods and services subgroups with those for 2011 and

24 Figure Miscellaneous goods and services, subgroup-level expenditure weights Micellaneous goods and services Subgroup level expenditure weights Personal care Personal effects Insurance Credit services Other miscellaneous services Source: Stats NZ Percent CPI all groups plus interest We excluded interest payments from the CPI since 1999, but have produced an analytical CPI all groups plus interest series since then. The interest component of this index has a September 2017 quarter weight of 7.34 percent, down from 7.37 percent in the June 2014 quarter. Mortgage interest has a September 2017 quarter weight of 5.83 percent, up from 5.62 percent in the June 2014 quarter. The main source of information for interest expenditure weights was the money, credit, and financial statistics published by the Reserve Bank of New Zealand. Our household living-cost price indexes (HLPI) also include interest. Tradables/non-tradables Tradables are goods and services that are imported or that are in competition with foreign goods and services, either in domestic or foreign markets. Non-tradables are goods and services that do not face foreign competition. Table 4 shows which of the items we added to the basket of goods and services this year are tradable and non-tradable. 24

25 Table 4 4 Tradable and non-tradable items added to the CPI basket Tradable and non-tradable items added to the CPI basket Tradable Fresh herbs Olives Flavoured tea Food preparations, mixers, and blenders Bicycle helmets Cellphone cases Headsets/phone sets Pet insecticides Source: Stats NZ Non-tradable Hem of trousers Storage costs Hearing aids Physiotherapist fees Admission charges to zoos Private accommodation rented from others Body massages The September 2017 quarter expenditure weight of the tradables component is percent, compared with percent in The September 2017 quarter expenditure weight of the nontradables component is percent, compared with percent in See table 6 in Consumers price index review: 2017 tradables/non-tradables CPI weights. tables 25

26 CPI pricing centres and regional expenditure weights This chapter describes the structure of our regional pricing centres and the regional expenditure weights. CPI prices are collected from 12 regional pricing centres within five broad regions (Auckland, Wellington, rest of North Island, Canterbury, and rest of South Island). We use regional weights to ensure price changes in a region with a larger population (eg Auckland) will have a greater effect on the national CPI than price changes in a region with a smaller population (eg Wellington). Instead of using the population weight of a region to determine its regional weight in the CPI, we base our regional weights on regional household spending from the HES. Regional pricing centres The 12 regional pricing centres are: Whangarei, Auckland, Hamilton, Tauranga, Napier-Hastings, New Plymouth, Palmerston North, Wellington, Nelson, Christchurch, Dunedin, and Invercargill. Regional expenditure weights versus population weights price change, and regional spending patterns to weight the individual basket items within each region. In contrast, regi regional price change and national spending patterns to weight the individual basket items in each region. Regional expenditure weights give a higher weight to regions with more spending per person compared with regional population weights. For example, percent of household expenditure is from households in Auckland, which has percent of the population. This means that, on average, price changes in Auckland will have a greater influence on the CPI using regional expenditure weights than regional population weights. Regional expenditure weights also give higher weights to items within a region where relative expenditure is higher than the national average. For example, housing and household utilities has a national expenditure weight of percent, but makes up percent of household expenditure in Auckland. This means price change for housing and household utilities in Auckland has more influence on the Auckland index using regional expenditure weights, than regional population weights, which use the national average weight. This produces more accurate regional indexes. Calculating regional expenditure weights We calculated regional expenditure weights as proportions of national expenditure for each CPI class or section (the lowest published level) using HES regional expenditure. We applied class/section level proportions to the individual items within that class or section (eg the regional proportions for fruit were applied to national expenditure on each fruit item) to derive regional expenditure on each individual item (eg spending on apples in Auckland). 26

27 In some cases, HES data was not available, and/or household sample sizes were too small to provide reliable regional breakdowns. Where this was the case we used data from other sources, or population shares. More than three-quarters of the regional expenditure weight was derived from the HES, while percent came from other sources, and the remaining 8.09 percent used population shares. Regional expenditure was then expressed in September 2017 quarter prices for the respective region (eg apple expenditure in Auckland was expressed in September 2017 quarter apple prices collected in Auckland). For broad regions with multiple pricing centres (rest of North Island and rest of South Island), we used population shares to allocate the regional expenditure weight to the pricing centres. Figure 13 e population. Figure CPI regional expenditure and population proportions Percent Auckland Wellington Rest of North Island Canterbury Rest of South Island Source: Statistics New Zealand CPI regional expenditure and population proportions 2017 CPI weight Population See tables 3 and 4 in Consumers price index review: 2017 tables revised 27

28 Reweighting the basket This chapter explains how we reweighted the CPI basket. Background to reweighting We reweight the CPI every three years, on average, as part of regularly scheduled CPI reviews. Reweighting ensures that the relative importance (expenditure weights) of the goods and services in the CPI basket continue to reflect up-to-date household spending patterns. The frequency of reweights is within the International Labour Organization (ILO) recommendation of at least once every five years. The 2017 reweight was based on data from the 2015/16 HES and other sources. The previous reweight, completed in 2014, was based on the 2012/13 HES and other sources. Data sources for reweighting the CPI basket The HES is the primary information source for reweighting the basket. However, other sources are also required, as the HES does not provide accurate expenditure estimates for some goods and services. HES respondents tend to under-report expenditure on some goods and services (eg tobacco and alcohol). Large, infrequent purchases (eg new cars) may not be reported frequently enough by the 3,500 households in the survey to provide accurate estimates of total household expenditure. We complemented the HES data with information from other sources, including Stats NZ surveys, government administration data, retail transaction data, and information provided by businesses. Over the last two CPI reviews (2014 and 2017), we changed our approach to make use of more HES data, when estimates from other sources are close to the HES estimates. In the 2011 review, 54.5 percent of the CPI weights were derived from the HES. This increased to 70.8 percent in In the 2017 CPI review, 66.8 percent of the CPI weights were derived from the HES. What the basket represents The goods and services in the CPI basket are a sample that represents the wider range of goods and services households buy. The expenditure weights assigned to the 701 individual goods and services in the new 2017 basket represent expenditure on those goods and services. In addition, the expenditure weights also represent expenditure on similar goods and services not directly included in the basket but expected to experience similar price change. For example, oranges and mandarins are in the basket but lemons are not. Expenditure on lemons (and other citrus fruit not in the basket) is allocated to oranges and to mandarins in proportion to spending directly on each of these two citrus fruits. Excluding out-of-scope expenditure When we use other data sources for expenditure estimates, we often have to adjust the data so 28

29 The HES (and CPI) reference population is New Zealand-resident, private households living in permanent private dwellings. This means that the reference population does not include: overseas visitors who expect to be resident in New Zealand for fewer than 12 months people living in non-private dwellings such as hotels, motels, boarding houses, hostels, motor camps, and homes for the elderly which provide medical care and services patients in hospitals residents of psychiatric and penal institutions; members of the permanent armed forces members of the non-new Zealand armed forces overseas diplomats. Children, including those at boarding schools, are not surveyed in the HES, but expenditure on behalf of those children by their parent or guardian is included. spend four or more nights a week in the household. They must share food consumption or contribute some portion of income towards providing the essentials for living as a group. How we exclude out-of-scope data Because the HES and CPI reference populations are the same, we do not need to adjust HES expenditure when using this source to estimate expenditure weights. However, a key step when we use other data sources to estimate expenditure weights is to adjust our estimates so they include expenditure made only by the HES/CPI reference population. If we sourced from the HES. In addition to the people listed above who are not covered by HES and CPI, we also exclude expenditure by businesses and government. We source other information from a wide range of providers. We asked data suppliers to report expenditure information that closely matches the scope of the CPI where possible. However, many providers were not able to break their data down to this level of detail, and instead provide data with a wider scope. To account for this, we estimate what proportion of a particular expenditure was in scope of the CPI. We derived a set of specific adjustment ratios for various areas of the basket to exclude out-ofscope expenditure. Appropriate ratios were then applied to the corresponding independently sourced expenditures. For example, these ratios took into account the relative share of expenditure by visitors from overseas, which varied significantly for different parts of the basket. Table 5 shows some of the information sources used to derive the out-of-scope adjustment factors. 29

30 Table 5 4 Information sources for out-of-scope adjustment factors 5 Information sources for out-of-scope adjustment factors Information sources for out-of-scope adjustment factors Type of out-of-scope expenditure Business and government Non-private and nonpermanent households Non-residents (overseas visitors) Source: Stats NZ Information source National accounts household consumption expenditure Breakdown Estimates proportion of total retail spending by the business sector, by retail store type 2013 Census Resident population Tourism satellite account International visitor expenditure as a proportion of total supply, by product Price updating expenditure the 2015/16 HES (although expenditure on some goods and services is collected in the HES on a one-year recall basis, meaning that purchases can span the two-year period from July 2014 to June 2016). When we used other sources to derive expenditure weights, we used information for the year to June 2016 where possible. by the ILO and is common international practice. The effect of price updating is to express the quantities underlying the 2015/16 expenditure values in the prices of the September 2017 quarter, the new price reference period. In general we used lower-level CPI indexes that correspond to the goods and services in the new basket to price-update the expenditure weights. Effect of price updating After price updating to the September 2017 quarter, the new CPI expenditure weights were about 2.75 percent higher. Price updating increased expenditure for percent of the basket, by weight. Prices for dwelling insurance increased by about percent from 2015/16 to the September 2017 quarter. Prices for cigarettes and tobacco increased percent over the same period. Price updating decreased expenditure for percent of the basket, and the remaining 0.43 percent experienced no price change. Goods and services where price updating resulted in reduced expenditure included: games and toys (down percent), international air fares (down percent), and accommodation services (down 7.50 percent). We customised price updating indicators for some items, such as new vehicles, telecommunications equipment, and audio-visual equipment and computing equipment, to include rather than exclude quality improvements. In general, this reduced the downward effect of price updating on the new weights, and in some cases, increased the 2015/16 weight. 30

31 Volume adjustments CPIs are generally calculated using a base-weighted Laspeyres-type (or Lowe) formula that measures the changing cost of acquiring a fixed basket of goods and services. The underlying are held fixed. We fix the underlying 2015/16 quantities in the Laspeyres-type price index formula over the life of the index (from 2017 to 2020). We assume that these quantities will be broadly representative of household purchases during the three-year life of the index (although we know that consumers will, to some extent, substitute towards goods and services showing lower relative price change during the period). We also know that there were significant shifts in quantities for some goods and services between the 2015/16 weight reference period and the September 2017 quarter price reference period. As part of the reweight, we made volume adjustments to some of the underlying quantities. This occurred selectively where there was strong evidence that a significant trend (rather than shortterm) change in volumes had taken place between the weight reference periods and the price reference periods. Although using selected adjustments introduces a level of subjectivity to the reweighting process, we consider this preferable to steadfastly retaining the underlying 2015/16 quantities across the whole basket. The aim of making these adjustments is to improve the relevance of the expenditure weights during the period they will be used ( ). before price updating, or incorporated into the indicator used to price update 2015/16 expenditures. Table 6 lists the volume adjustments we made, the reasons for making them, and the data sources used to determine whether to adjust and by how much. Note that volume change includes not only change in physical quantities, such as the number of new vehicles acquired, but also change in quality. Table 6 tems 6 Volume adjustments for CPI items Volume adjustments for CPI items Goods or service Reason for adjustment Information source Cigarettes and tobacco New cars Telecommunication equipment and audio visual equipment Source: Stats NZ To reflect lower volumes, influenced by excise tax increases To allow us to better calculate weights for newer models and discontinued models To reduce the downward effect of price updating due to quality adjustments Cigarettes and tobacco available for consumption data for the years to June 2016 and June 2017 Vehicle registration information for the years to June 2016 and June 2017 Volume and sales information available from GfK administrative data 31

32 Managing cyclical or volatile expenditure Expenditure for some goods and services is highly cyclical or highly volatile. Two highly-weighted examples are the purchase of newly constructed dwellings, and insurance services. Since the 2006 review we have extended the weight reference period to three years (ending the same date as the HES) for purchase of housing (and related services such as conveyancing and real estate fees), rentals for housing (for 2006, 2008, and 2011 only), and insurance services. This extension makes it possible to partially smooth the impact of cyclical highs or lows (housing) or unusual claims years (insurance). For the 2017 review, we used information for the three years to June In cases where we made volume adjustments, and for housing and insurance (where we used an average of more than one year for the weight reference period), we adjusted volumes for periods after or before the weight reference period of 2015/16 to remove the effect of growth in the usually resident population. For example, nominal growth in volume from the weight reference period of 2015/16 was deflated by population growth between 2015/16 and 2016/17. This action ensured any volume adjustments we made were above those that may have resulted from population change alone. Household expenditure levels The CPI Advisory Committee 2013 recommended we express the CPI weights as dollar values (average weekly household spending) as well as percentages. the CPI, the Explore living-costs in New Zealand app expresses the HLPI weights as dollar values and includes an all households index. 32