Stats NZ has a new website.

For new releases go to

As we transition to our new site, you'll still find some Stats NZ information here on this archive site.

  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+
Balance of payments income revisions and the international investment position

Statistics New Zealand introduced improvements to the coverage of the income component in the balance of payments (BoP) statistics in the June 2010 quarter. These changes are documented in Improvements to income in balance of payments statistics.

One of the improvements was the use of incomes data, as reported to Inland Revenue for tax purposes, to estimate investment income flows that were not previously measured as part of the BoP statistics. The inclusion of this additional net income in the BoP has raised further questions and speculation about the size of the stock of overseas assets and liabilities that give rise to these investment income flows. Westpac, in their preview of the June 2010 quarter balance of payments and international investment position (IIP) release, suggested that the additional income may imply net overseas asset levels that would significantly lower the ratio of the net international investment position (IIP) to gross domestic product (GDP) from -86.5 percent of GDP (as at June 2010 quarter) to something closer to -60 percent of GDP.

This article seeks to increase the understanding of this issue by clarifying which investment incomes and associated stocks of assets and liabilities are currently measured in New Zealand's BoP and IIP statistics and which are excluded.

Current coverage of BoP and IIP statistics

The balance of payments are a set of flow statistics that measure transactions between New Zealand residents and the rest of the world. The BoP includes transactions in goods, services, income flows from assets held overseas and liabilities incurred by non-residents, and transfers and financial transactions between residents and the rest of the world.

Prior to the latest revisions, the main measurement gap was the income flow (inflows and outflows) of individuals, smaller-sized companies, and other business type entities (such as partnerships, and trusts and estates). These small units were not captured in the existing survey and/or administrative data sources being used. However, this gap is, in principle, now covered by the use of income data from Inland Revenue – ‘in principle’, as any overseas income flows not declared to Inland Revenue by these small units will remain omitted.

It should also be noted that the retained earnings (debits or credits) of any direct investors1 included in the undeclared overseas income flows will also be excluded, contrary to BoP guidelines. The additional measured income will be confined to actual income flows declared for tax purposes. However, this omission is not thought to be very significant, given the types of economic units covered by the Inland Revenue-sourced incomes data.

International investment position statistics are a set of stock statistics measuring New Zealand's liabilities and assets with the rest of the world. The latest net IIP, as at 30 June 2010, was recorded at negative NZ$163.7 billion or -86.5 percent of GDP.

The IIP statistics are sourced from the same surveys that are used to measure the investment income flows in BoP statistics. However, there are some known gaps in the coverage of New Zealand's stock of international assets and liabilities. These include:

  • New Zealand investment abroad: overseas assets directly held by unmeasured New Zealand resident units such as individuals, smaller-sized companies, and other small business type entities (eg small managed funds, trusts and estates, and partnerships).
  • Foreign investment in New Zealand: liabilities held by unmeasured New Zealand resident units such as those described above.

It is generally assumed that the unmeasured units are more likely to hold assets (eg shares, overseas property, or funds in overseas bank accounts) than liabilities. While Statistics NZ makes annual estimates of (a) the direct equity share holding in overseas companies by New Zealand individuals; and (b) the overseas assets held by small funds managers, these are not considered sufficiently accurate to include in the published IIP statistics (the estimates are, however, published in the Technical notes to the Balance of Payments quarterly releases). For example, the estimation methodology for (a) relies on several key assumptions that cannot, at this stage, be fully tested by reference to available data. Changes to these assumptions can significantly alter the size of the estimate.

Aligning the two sets of statistics

Figure 1 illustrates the lack of alignment in the coverage of the two sets of statistics. As can be seen, the BoP coverage is, in principle, now complete, whereas that for the IIP is not. Furthermore, the survey coverage of the IIP is greater than the survey coverage of the matching BoP flows – this is shown by the dotted line in the table. What this indicates is that some of the assets and liabilities that give rise to the additional Inland Revenue-sourced income are already included in the IIP via Statistics NZ surveys.

Figure 1

Coverage of international liabilities and assets in the IIP and corresponding investment income in the BoP statistics

 Figure, Coverage of international liabilities and assets in the IIP and corresponding investment income in the BoP statistics.

Note: Not drawn to scale.

The composition of the Inland Revenue income flows

The additional investment income credits have been derived from tax returns for both individuals and estates (IR3, IR6, and IR7 returns), and companies (IR4 and IR4F returns). From a statistical perspective, one of the limitations in using this data is that the ‘overseas income’ is captured in a single field and it is not possible to distinguish between different income sources. Information included on an IR3 return, for example, could include income from labour, rentals, investments, royalties, service contracts, and even a pension. In the absence of other information, it has been assumed that most income will relate to investments and all of the income has been classified in the BoP accordingly.

This income identification problem does not apply to the additional investment income debits. These are sourced from the IR67 and IR3NR returns, which have separate fields for interest, dividends, and other distributions, where these are either remitted overseas (IR67) or earned by a non-resident (IR3NR).

The investment income credits data also suffer from a number of unmeasured quality issues:

  • True investment income is likely to be underestimated due to exemptions applied to overseas income from grey-list countries2 (removed in 2009) and the four-year exemption granted to recent migrants (from 2006).
  • Changes in the foreign investment fund regime may lead to switches in declaring income from a ‘fair dividend’ basis to one involving capital losses.
  • The latest year’s estimates are based on incomplete tax data and involve some extrapolation of previous year’s returns. Given the significant fall in asset values in 2008/9, returns in these earlier years will be a poor guide to the latest year’s income.

Estimating the unmeasured component of the IIP

From figure 1 it can be seen that it may be possible to estimate the unmeasured assets and liabilities in the IIP by working back from the investment income debits and credits they generate (now measured via the Inland Revenue-sourced income data) and assumed yields. Any values so derived are very sensitive to (a) the assumed rates of return and (b) the percentage of the additional BoP ‘investment income’ credits that are assumed to truly reflect investment income.

Allowance also needs to be made for the non-alignment of the BoP and IIP survey sources. In particular, the assets and liabilities of enterprises recorded in the IIP but not in the BoP need to be deducted before deriving additional IIP assets and liabilities using the above method otherwise they will be double counted.

Using the above approach, Statistics NZ has made an estimate of the unmeasured component of the IIP based on the following simple assumptions:

1. Investment income inflows. The income flows are first split into business and household income based on the type of tax return. It is then assumed that 66 percent of the business’ overseas income and 50 percent of the households’ overseas income is true investment income. These percentages are entirely hypothetical as, at this stage, no additional information is available to more accurately split the single ’overseas income’ by source.

2. Investment income outflows. These are split into interest and dividends as given in the Inland Revenue data.

3. Yields. Using unpublished data on BoP portfolio investments (debits and credits),3 average yields for the last five years (2006-10) were obtained. These were:

  • Assets: equity 2.09 percent and debt 7.77 percent (a simple average of 4.93 percent is used in the calculations as the IRD data does not provide an instrument split); and
  • Liabilities: equity 7.22 percent and debt 4.13 percent.

The result was a March 2010 net IIP estimate little different to that published in the release for the March 2010 quarter. The additional international assets (NZ$20.9 billion) were largely offset by additional liabilities (NZ$17.9 billion). The net IIP to GDP ratio changed from -85.9 percent to -84.3 percent. Of course, different assumptions would produce different results.

It must be emphasised that this estimate has been produced solely for illustrative purposes. While the assumptions are considered ’reasonable’, the resulting estimates should not be seen as official statistics.


Statistics NZ has been very cautious in including the Inland Revenue-derived income data into the BoP. While the data contains a number of weaknesses that are difficult to quantify, it is nevertheless considered a good proxy for income flows that were not previously captured. This BoP revision is consistent with the Statistics NZ data collection strategy which will lead increasingly to data being primarily sourced from administrative sources as opposed to direct collections.

However, the new data has not been used to model estimates that might be used to close a gap in the coverage of the IIP statistics. While simple estimates can be derived using the approach described above, it is considered that further work is needed on improving the breakdowns of the incomes data and on verifying the yield assumptions before such estimates might be included in the official statistics.


1. A direct investor is either an individual or an enterprise which owns 10 percent or more of the ordinary shares or voting power (for an incorporated enterprise), or the equivalent (for an unincorporated enterprise), in an enterprise located in an economy other than the one in which the direct investor resides.
2. The grey list countries were Australia, Canada, Germany, Japan, Norway, Spain, the United Kingdom, and the United States.
3. The unmeasured units in the IIP are mainly individuals or small business units and it was considered that their investments would be predominantly portfolio in nature.


Salendra Kumar at Statistics New Zealand (


Previously published as: Cope, J & Kumar, S (2010, November). Balance of payments income revisions and the international investment position. Asymmetric Information, 39, 10–11.

December 2010

  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+
  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+