Brian Fallow: Kiwis may not be such poor savers after all

A select group are being quizzed every other year about their assets and liabilities. Photo / Glenn Jeffrey

Research by Treasury economists casts doubt on the conventional view that we are hopelessly improvident and feckless.

The conventional view is founded on estimates of household saving rates which show households year after year spending more than they earn - $1.14 for every $1 of income in the year to March 2009.

But saving rates are slippery things to grab hold of and measure.

In particular - when you are looking at the difference between two very big big numbers, such as households' collective incomes and consumption spending, or their collective assets and liabilities - all the errors in calculating those numbers accumulate in the difference between them.

The most you can hope for is that the errors will be consistent over time, so that you can at least get a handle on the trend.

What is valuable about the work by the Treasury's Grant Scobie and Katherine Henderson is that it is based on a new source of data called Sofie (Statistics New Zealand's survey of family income and employment).

Instead of an aggregate or top-down perspective based on the national accounts, it provides a bottom-up view through extensive interviews with a large number of people, around 16,000.

Crucially, they are the same people each year, which avoids some limitations of other survey-based data such as the household economic survey.

Every other year they are quizzed about their assets and liabilities. This allows Scobie and Henderson to use one of the ways of measuring savings, which looks at changes in people's net wealth over time.

This is the "stock" approach as distinct from the "flow" approach, which subtracts what people spend on consumption from their income and calls the balance saving (or in New Zealand's case dis-saving, since the estimates, based on top-down national data, indicate we consistently spend more than we earn).

By comparing the state of people's balance sheets in 2004 and 2006 (the most recent year for which the data have been made available) the researchers can see how much people's net wealth increased over that period.

That is then divided by their disposable (or after-tax) incomes to derive an annual saving rate.

It turns out to be 41 per cent.

Not bad, you might think.

But that is in nominal terms. The most obvious immediate problem is the period between 2004 and 2006 falls within the most recent housing market boom, when house prices soared.

"Passive saving in the form of revaluation of house prices constituted a major part of the total change in net wealth," Scobie and Henderson say.

When they adjust for house price inflation over that period, the saving rate drops to 12 per cent on average.

That is still a very different picture from the minus 10 per cent recorded by the flows measure (based on national data for household incomes and outlays) over the same period.

But it is in line with earlier work by the Reserve Bank on household balance sheets.

There is, however, an argument about whether housing should be included in calculating net wealth at all. If the aim is to measure the total stock of household wealth then it is legitimate to include changes in the valuation of all assets including housing, say Scobie and Henderson.

"On the other hand if one is interested in retirement saving and one is prepared to start from the assumption that many retirees remain in their pre-retirement home, or even if changing homes buy a smaller, newer one of equivalent value, then one might well wish to exclude the value of the principal residence."

When they exclude owner-occupied housing from the figures, the median saving rate drops from 16 to 5 per cent - lower, but still positive, in contrast to the persistently negative saving rates generated by the flow method.

Scobie and Henderson offer a couple of reasons for being wary of the flow approach. One is that it is derived from the national accounts, which treat spending on education as consumption rather than investment.

But a similar problem exists with the stock or balance-sheet approach. Student debt is recognised as a liability, but there is no corresponding asset reflecting the human capital (hopefully) acquired in the course of running up that debt.

Yet university graduates make up around 17 per cent of the working age (over 15) population, but nearly 40 per cent of the top income decile (those earning more than $75,000 a year).

Another problem Scobie and Henderson point to with the flow approach is that it has been consistently negative since 1993.

If that were accurate, the cumulative effect by now would be that households' net wealth was negative - "a clearly implausible result".

While this study presents a more cheerful picture of household saving than the flow-based conventional view, at this stage it is only a snapshot.

It shows households were saving between 2004 and 2006, but that gives us no information about the trend. Future "waves" of Sofie data should cast some light on that. Nor does it address the issue of whether people are saving enough for their old age - however one might define "enough".

After all, in the course of a lifetime there are times when it makes sense to run up debt, times when it makes sense to accumulate assets and times when it make sense to run savings down. Aggregate or median numbers shed no light on that.

A puzzling feature of the data Scobie and Henderson report is that however they slice and dice the survey population - by income decile, by age group, by labour-market status or by health status - there is a consistent pattern of around 40 per cent of those sub-groups who have negative savings rates.

There is some variation along the lines you would expect: more negative savers in the lowest income decile than the highest, more among the youngest and oldest age groups than those in between, and more among those in poor health than good health.

But what is striking is they are clustered within a few percentage points of 40 per cent across such a wide range of individual circumstances - leaving the researchers scratching their heads.

While this early harvest of the Sofie data is encouraging and intriguing, it would be premature to break into a full-throated rendition of the Hallelujah Chorus.