What NZ Households Actually Spend, and How to Read It
📊 Where the numbers come from
Almost every claim you will read about what New Zealanders spend traces back to one source: the Household Economic Survey, run by Stats NZ. It asks a sample of households to record what they spent, and it publishes the results as average weekly household expenditure across about seventy-eight categories. It runs every three years. The most recent is the 2023 survey, and the ones before it were 2019, 2016, 2013, 2010 and 2007.
Add the thirteen top-level categories together and the average New Zealand household spent about $1,597.50 a week in the 2023 survey. That figure is genuinely useful and it is also easy to misuse, in four specific ways. This guide is about those four, because a spending figure read carelessly is worse than no figure at all: it looks precise, it feels authoritative, and it can tell somebody they are failing when they are not.
One thing to get straight immediately: the survey measures outgoings, not consumption. It counts money the household put into savings and investments as expenditure, because from the household's point of view the money left. So the total above is not what New Zealanders consume. It is what leaves the account.
You can see every category, with its own history back to 2007, on the household spending pages. Each category page breaks the figure down by household size, region, tenure and income decile, and each sub-category has its own page explaining what it does and does not include.
⚖️ An average is not a typical household
The survey publishes averages. An average is the total divided by the number of households, which means a small number of very high spenders pull it upward. The median, the figure that half of households sit below, is almost always lower. The survey does not publish medians for these categories, so every number you see on these pages is an average and should be read as one.
This matters most where spending is concentrated in a minority of households. The survey publishes, alongside each figure, the share of surveyed households that recorded any spending in that category at all. That percentage is the single most useful column on the page, and it is the one almost nobody looks at.
Alcohol is the clearest example. About 35 per cent of households recorded any alcohol spending in the survey period. The all-household average is therefore roughly a third of what a household that does buy alcohol actually spends, and the gap is not a rounding matter. Read the average as a national total divided up, not as a description of anyone.
Grocery food sits at the opposite end. Over 94 per cent of households recorded grocery spending, so the average of about $139.40 a week is close to a real household's bill. The same reasoning that makes the alcohol average misleading makes the grocery average trustworthy.
🏠 The rent example, which catches almost everyone
Averaged across every household in New Zealand, spending on rent was about $144.80 a week in the 2023 survey. That number is widely quoted and it is close to useless, because most households do not rent. Only about 34 per cent recorded paying any rent at all.
Restrict the same survey to households that actually make rent payments and the figure is about $448.80 a week, recorded by nearly 99 per cent of them. That is the number a renter should compare against. The difference between $144.80 and $448.80 is not a different survey or a different year. It is the same data asked a better question.
The equivalent for owners is split across two categories, which is its own trap and is covered in the next section.
🏷️ A category rarely means what it sounds like
The survey uses a formal classification, and its boundaries are not the ones a household would draw. A few that reliably catch people:
- Milk is not a beverage. It is classified as a grocery food. Non-alcoholic beverages means coffee, tea, soft drinks and juice.
- A cafe coffee is a restaurant meal. Anything bought ready to eat or drink sits in restaurant meals and ready-to-eat food, so a household that spends heavily on takeaway coffee shows almost nothing under beverages.
- Mortgage interest is not a housing cost. It is recorded under other expenditure, in a category of its own, separate from the home ownership costs recorded under housing. For owners who make mortgage payments the two are of similar size, about $330.80 and $315.30 a week respectively, so using either one alone understates the cost of owning by roughly half.
- Insurance is not a housing cost either. House, contents, vehicle and health insurance are all recorded together under miscellaneous goods and services.
- Nappies are personal care, not a household supply and not a food.
- Saving is expenditure. Contributions to savings and investments are recorded as money going out, which is why the survey total is not a consumption figure.
None of this is the survey being difficult. A classification has to put every dollar somewhere and the boundaries have to be consistent across decades. But it does mean that comparing your own budget line against a survey category requires knowing which category your line actually falls into. Every sub-category page on this site opens by saying what it covers and, more usefully, what it excludes.
📏 Sampling error, and when a number is not safe to use
The survey asks a sample of households, not all of them, so every figure carries a margin. Stats NZ publishes that margin with each number, as a percentage of the figure itself. A sampling error of 3 per cent on household energy means the true average is very likely within a few per cent of the number shown. A sampling error of 40 per cent means the figure could plausibly be a third out in either direction.
On this site, any figure whose sampling error exceeds 30 per cent is not shown as a number. The page says so instead, and explains why. That threshold is a judgement, not a rule from Stats NZ, and it is applied consistently: a precise-looking number that could be a third out is worse than an honest gap, because a reader will act on the first and ask questions about the second.
💵 Old money and today's money
The survey runs every three years, so comparing 2007 with 2023 compares two different dollars. A category that appears to have risen 80 per cent may have risen very little in real terms once inflation is removed. Where it is possible, the pages on this site show both: the figure as published, and the same figure restated in current dollars using the matching Consumers Price Index series. The restated column is arithmetic done here rather than a Stats NZ statistic, and it is always labelled that way.
Two of the thirteen categories cannot be restated at all, because the CPI has no series that matches them. Other expenditure and sales, trade-ins and refunds are not purchases of goods and services, so there is no price index for them. Those pages say so rather than showing an empty column.
🎯 Using it on your own budget
The point of a benchmark is not to tell you whether you are good with money. It is to tell you which of your costs is unusual, so you know where to look. Most households have one or two lines well outside the normal range and several that are unremarkable, and only the outliers are worth your time.
Three rules make the comparison honest:
Compare against your own household, not the country. A one-person household spent about $132.40 a week on food in the 2023 survey, and a four-person household about $437.30. Using the all-household average of $299.50 tells a single person they are frugal and a family of four they are extravagant, and neither is true. The same applies to tenure for housing, and to region for most things.
Match the period. The survey publishes weekly figures. If you budget monthly, multiply by 52 and divide by 12, which is about 4.33, not 4. Over a year that difference is about a month of spending, and it is the most common arithmetic mistake made with this data.
Do not add breakdowns together. The survey publishes spending by household size, and separately by region, and separately by tenure. It does not publish a figure for two people in Wellington who rent, and combining the three breakdowns to invent one produces a number the survey never supported. Use whichever single breakdown is most relevant to the question you are asking.
The Budget Calculator does this comparison for you. Enter your own figures, choose your household size, region and tenure, and it shows each category beside the survey average for households like yours, converting to whatever period you are budgeting in. It benchmarks housing against the right tenure rather than a national average, which is the difference between a useful comparison and an alarming one.
🧭 What the data cannot tell you
It cannot tell you what you should spend. A survey describes what households did, not what is wise, and the average includes households in serious difficulty alongside households with plenty of room. Spending less than average on food is not automatically good and spending more is not automatically bad.
It cannot separate prices from behaviour. When a category rises between surveys, some of that is things costing more and some is households buying different things, and the survey has no way to tell you which. That is why the inflation-adjusted column matters and why even it is only half an answer.
It cannot capture what people would rather not report. Alcohol and tobacco are understated in household surveys everywhere, and the error runs in one direction only.
And it is not current. The 2023 survey is the most recent, and prices have moved since. Use it for structure and proportion, which change slowly, rather than for the exact price of anything, which does not.
🎯 Test Your Knowledge
Quiz on reading New Zealand household spending data
Frequently Asked Questions
How much does the average New Zealand household spend each week?
Adding the thirteen top-level categories of the 2023 Household Economic Survey gives about $1,597.50 a week. It is an average rather than a typical household, and it includes money put into savings as well as money consumed.
Is the average the same as what a typical household spends?
No. An average is pulled upward by a small number of households spending a great deal, so the typical household usually spends less. The survey publishes averages, not medians.
Why does the average rent figure look so low?
Because it is averaged over every household, including the majority who own and pay no rent. Restricted to households that actually make rent payments, the 2023 figure is about $448.80 a week rather than $144.80.
What does sampling error mean on these figures?
It is the uncertainty that comes from surveying a sample rather than every household, published as a percentage of the figure. A large sampling error means the true value could sit well above or below the number shown, which is why figures above 30 per cent error are not shown as numbers on this site.
How often is the Household Economic Survey run?
Every three years. The most recent is 2023, and the surveys before it were 2019, 2016, 2013, 2010 and 2007.
Related guides and data
- Household spending data, every category with its own history back to 2007.
- Budget Calculator, which compares your own figures against households like yours.
- Budgeting Methods for NZ Households, on choosing an approach you will stick to.
- Needs, Wants and the Line Between, the judgement a benchmark cannot make for you.
- Sinking Funds, for the lumpy categories that ruin a monthly budget.