Saving Rates and Portfolio Allocation in New Zealand Joint Working Group Government Officials and ISI

Joint Working Group, Government Officials and the Investment Savings and Insurance Association (ISI) Working Paper Saving Rates and Portfolio Allocation in New Zealand Joint Working Group Government Officials and ISI New Zealand Treasury Working Paper, No. 99/09 Provided in Cooperation with: The Treasury, New Zealand Government Suggested Citation: Joint Working Group, Government Officials and the Investment Savings and Insurance Association (ISI) (1999) : Saving Rates and Portfolio Allocation in New Zealand Joint Working Group Government Officials and ISI, New Zealand Treasury Working Paper, No. 99/09, New Zealand Government, The Treasury, Wellington This Version is available at: https://hdl.handle.net/10419/205412 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Disclaimer: This paper has been prepared by a joint working group comprising Treasury, the Inland Revenue Department, the Reserve Bank of New Zealand and the Investment Savings and Insurance Association. The views expressed are a preliminary group consensus and may not fully express the views of any individual organisation. The Treasury takes no responsibility for any errors or omissions in, or for the correctness of, the information contained in these working papers. TREASURY WORKING PAPER 99/9 Saving Rates and Portfolio Allocation in New Zealand Joint Working Group Government officials and ISI ABSTRACT This paper explores readily available data up to 1996 on New Zealand’s saving rates and the level of savings, and examines the portfolio allocation of savings. It finds that New Zealanders’ household saving rates are low, whilst above average public saving rates leave national saving rates closer to OECD average levels. Whilst New Zealand has lower net wealth per household than many of its OECD counterparts, a good portion of this is associated with differences in per capita incomes. New Zealanders hold a higher proportion of their savings in housing than people in many other countries, as would be expected given our lower per capita incomes. Even given this, New Zealand has above average home ownership rates and housing space consumption is high by world standards. Historically, the returns on different domestic asset classes suggest that the risk adjusted real return on housing has been relatively high. 1. INTRODUCTION ......................................................................................................................2 A)B ACKGROUND ..........................................................................................................................2 B)M EASURES OF THE LEVEL, MIX AND RETURNS TO DIFFERENT TYPES OF SAVING IN NZ..................3 i) Measurement of the Level of Saving.....................................................................................3 2. NEW ZEALAND’S SAVING RECORD....................................................................................5 3. INTERNATIONAL COMPARISON OF SAVING LEVELS...................................................7 A)C OMPARISON OF NEW ZEALAND HOUSING INVESTMENT WITH OTHER COUNTRIES......................9 4. TYPES OF INVESTMENTS HELD BY NEW ZEALAND HOUSEHOLDS......................... 12 A)R ETURNS................................................................................................................................13 i) Adjustments to the Housing Series...................................................................................... 13 ii) Cumulative Returns for Different Time Periods..................................................................16 5. THE IMPACT OF TAXATION ON INVESTMENT DECISIONS........................................ 18 A)L EVERAGING INVESTMENTS .................................................................................................... 21 6. FURTHER RESEARCH.......................................................................................................... 22 7. CONCLUSIONS....................................................................................................................... 23 APPENDIX: POTENTIAL FACTORS INFLUENCING INVESTMENT AND SAVING DECISIONS ...................... 24 a) Demographic Factors........................................................................................................24 b) Per Capita Income............................................................................................................. 24 c) Credit Institutions.............................................................................................................. 25 d) Social Security Programmes .............................................................................................. 25 e) Other Government Actions................................................................................................. 25 f) Returns to Schooling..........................................................................................................26 REFERENCES....................................................................................................................................28 2 1. INTRODUCTION In September 1998 Treasury commenced a project with the Investment Savings and Insurance Association (ISI) to study the data available on New Zealand levels of saving and patterns of investment. The aim of this work is to identify and fill gaps in our knowledge in order to assist in future policy development. The project was planned to be undertaken in the following manner: Phase 1 : Collection of data from pre-identified sources. Phase 2 : Initial analysis of data collected in Phase 1. Phase 3 : Part A – detailed analysis of the data. Part B – improving the data. The current document is the result of phase two of this saving research project. Central questions in this stage are: What patterns have emerged from the data? What looks interesting/unusual and therefore warrants further work? What gaps are there in the data that we would like to fill? a) Background In June of 1998 the Investment Savings and Insurance Association (ISI) published a report aimed at contributing to the debate on the appropriate retirement policy for the future. As a result of the report, the Minister directed officials to research some central questions from the ISI report on retirement saving “A Wakeup Call” [1998]. The particular issues of concern were: “6.3 Factors influencing the quality of investment and saving decisions Much of the analysis of retirement income issues to date has focused on the adequacy of savings from a quantitative point of view, rather than at the quality of saving and investment decisions which offer returns that reward savers for foregoing immediate consumption. More research could usefully be undertaken on the distribution of saving and investment decisions which offer returns that reward savers for foregoing immediate consumption. 6.4 The role of housing as a vehicle for retirement saving New Zealanders have historically had a strong preference for home ownership as a means of retirement saving. The debate on the merits of this preference has focused on whether the regulatory, taxation and legal frameworks encourage an over-investment in home ownership. 3 There would be considerable value added to this debate by looking at the reasons why individuals seek home ownership and what role this should ideally play in a retirement savings strategy. Regard must be given to the risks of having residential property as the primary asset in a savings portfolio. It would also be useful to compare a typical New Zealand savings portfolio with the optimal portfolio choice.” This paper is structured around two questions: does New Zealand save enough, and is this wealth invested optimally? The remainder of this paper commences with a brief description of saving. Following this we summarise historical saving behaviour for New Zealand and then compare this to other OECD countries. The composition of New Zealand’s portfolio of assets is then presented followed by the historical returns that four main asset classes have experienced. Some factors which have typically been suggested as affecting saving rates have been summarised in the Appendix. b) Measures of the level, mix and returns to different types of saving in NZ i) Measurement of the Level of Saving “Saving” is a term that is commonly used in several ways, e.g. (a) a household’s net worth, or (b) a change in its net worth, or (c) national aggregates of these two. Economists use the term “saving” to mean the annual flow measure of saving and “savings” are the stock measure. We attempt to follow this convention below. Aggregate household saving is measured as a residual of household income minus outlays, from the System of National Accounts household income/outlay account. There are several problems surrounding this measurement of saving, not the least of which is the fact that it is usually computed as the residual of two flows. Since this residual is typically much smaller than either income or outlays, its measurement is prone to large errors. An alternative measure of annual national saving to that provided in the System of National Accounts could be measured as the growth in the country’s stock of net worth from one year to the next. This would accurately include capital gains on housing1 and shares, which are excluded from the flows method. The disadvantage is that stocks of assets and liabilities (of which net worth consists) are not commonly measured in New Zealand’s official statistics. Hence studies usually rely on the less accurate flows approach which is likely to underestimate saving given that there is likely to be some positive level of capital gains. 1 Saving is the portion of income that is not consumed. This amount can then be invested into housing. In this manner, paying principal on a mortgage is an investment decision involving the reallocation of saving rather than saving itself. The portion of a housing investment which does represent saving is any capital gains that the household does not offset with increased expenditure as these will represent an increase in the wealth of the investor. 4 For the purpose of analysis, it would be preferable if statistics on saving rates by age groups were available. Aggregate measures reflect the saving or dissaving of generations approaching or already in retirement. By comparing differences between the saving rates of different age groups we could evaluate how changing factors are influencing saving decisions for different demographic groupings. 5 2. NEW ZEALAND’S SAVING RECORD National annual saving flows in New Zealand can be broken down into contributions by sectors such as households, business and Government. The statistics that are available demonstrate that the patterns of these types of saving over recent years have been quite different. As can be seen in Figure 1 below, there has been a decline in the level of annual household saving since the latter half of the 1980s, and to a lesser extent the early 1990s. It is interesting to note how this trend compares with the trend in Government sector saving. During the 1994–1996 period when the annual household saving fraction was at its lowest it can be seen that Government saving was significantly larger. The net result is that total national saving as a percentage of GDP is relatively constant during the 1994–1996 period, despite the large fluctuations in the constituent series. National saving did decline significantly during the early 1990s, as both household and Government saving declined. This is likely to be at least partly a business cycle effect, as New Zealand experienced a recession during that period. Figure 1: National Saving and its Components % of GDP 1987/88 1989/90 1991/92 1993/94 1995/96 12 12 10 10 8 8 6 6 4 4 2 2 0 0 -2 -2 -4 -4 Household Business Government National Source: Statistics New Zealand Note that only the Government saving statistic is considered to be an official statistic. Explanations offered for these movements in saving rates include: • The recent increase in house values has raised households’ net worth. As a result, households do not save as much, because they do not need to in order to achieve a given future level of consumption. These effects are probably stronger in Auckland than elsewhere. Note that capital gains (such as those resulting from increased house prices) are not included in 6 the flow measure of household saving. Hence, the actual total level of household saving is higher than the statistics indicate. • A Ricardian Equivalence argument: higher government saving leads people to expect that tax rates will be lower in the future, thus encouraging them to save less in the current period. Also, when people perceive the government to be saving by spending less on education and health care, individuals respond by spending more on those items themselves, resulting in lower measured household saving, though, of course, human capital may be accumulated. Sebastian Edwards [1995] finds that while government saving depresses private saving, this does not happen at a one to one rate: an increase in government saving of one per cent is associated with a decline in private saving of about 0.55 per cent, causing national saving to rise by 0.45 per cent. A stable political and institutional environment is found to be a strong promoter of government saving. Edwards notes that the less than perfect negative correlation between government and private saving can be explained by the fact that government saving is affected by political stability whereas private saving responds to demographic variables and social security expenditures. Both are affected by real growth and by the current account balance or foreign saving. It should be noted that, for an open economy with good access to foreign capital, it is not obvious that there is an optimal level of national savings, as any lack in domestic savings can be covered by an inflow of foreign savings. This, however, could become significant if the foreign borrowings are not invested for a sufficient return and the level of foreign debt becomes unsustainable. 7 3. INTERNATIONAL COMPARISON OF SAVING LEVELS All studies considered in the course of this project conclude that the gross level of annual national saving in New Zealand is at a level comparable to that of other OECD countries. International comparisons, however, are complicated by the fact that not only is household saving difficult to measure, there are definitional differences among countries as well. As a result, comparisons can be only indicative. Asset prices, both individually and collectively, behave very differently between countries2. This affects the ratios of individual asset values to the total portfolio in each country. Each type of asset value behaves differently from other assets in the course of a country’s business cycle. Not only that, there seems to be very little correlation between the prices of broad classes of assets between countries. We do not have rates of return data by asset for different countries. These being the facts, comparisons of, say, housing wealth as a fraction of households’ wealth or of GDP between countries are limited in what they reveal about the efficiency of the household portfolio in a particular country. Dean, Durand, Fallon, and Hoeller [1990] from the OECD, report that the correlation between national saving and national investment has generally become much looser since 1980 due to the liberalisation of financial markets, and resulting much larger international capital flows. Also, most OECD countries have seen a significant fall in national saving and gross national investment, with the fall in saving exceeding the fall in investment. In Dean et al., the level of national saving in New Zealand does not look out of the ordinary compared with that in other OECD countries. The most thorough international comparison relating to saving in New Zealand can be found in Savage [1997a]. Savage notes that net household annual saving as a proportion of disposable income in New Zealand is low by comparison with other OECD countries, particularly during the early 1990s. Net national saving as a proportion of GDP, however, which includes the Government and business sectors, is comparable to other OECD countries particularly during the early 1990s when Government saving was strong. Savage notes that the main difference between the flow and the stock method (defined on page 2) for measuring saving is that the flow method omits the capital gains from housing and shares. As this could be significant, the following graph has been constructed using the stock method. The OECD has compiled relative net worth measures (as a percentage of disposable household income) for some major OECD countries. These data, complemented by numbers from Statistics NZ and the Westpac-FPG household financial data3 for New Zealand, produced Figure 2. Again, due to difficulties with the data these comparisons should only be treated as indicative. 2 Borio, C, Kennedy, N and Prowse 1994 3 Data compiled by Westpac.FPG from surveys of deposit institutions. 14 home, the series would only catch the resulting increase in house price. This increase would be overstated as there was an additional cost of $20,000. We have tentatively estimated the value of home improvements at around 2% per annum. Improvements in the quality of new houses also affects the index and should be adjusted for. Taking these various factors into consideration we have adjusted the nominal returns to housing series upwards by 2.5% per annum. Note that this adjustment is based as much on estimates as it is on data and the reader should keep this in mind when interpreting the preliminary results presented below. The data on real returns for the four assets, including the housing adjustment, is summarised in the graph below. Figure 5: Annual Real Returns to Different Assets Source: Westpac FPG Note: These returns are not adjusted for risk. As can be seen from the graph, the stock market has been an unreliable (risky) performer, short and long term interest rates have been relatively stable, and house prices have mostly been somewhere in between. Phase three of this project will include a risk adjustment of the asset returns. Note that during the high inflation, interest rate control period of the 1970s real returns from fixed interest investments, and to a lesser extent from housing, were negative. Table 3 displays the average annual real return and the effective annualised real return for the whole period for the four assets. Annual Real Returns -60 -40 -20 0 20 40 60 80 100 120 140 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 Percent nzse40 annual real real 6 month deposit rate real 10yr bond real housing with adjustment 15 Table 3: Returns for New Zealand Asset Classes Asset Average Annual Real Return (19701998) % Effective Annualised Real Return for 197019988 (%) NZSE40 9.97 5.54 Housing 4.83 4.38 6 month deposit rate 0.86 0.72 10 year Government Bonds 1.36 1.23 One point to note is that even if average annual returns from shares are larger than that for housing it is not immediately obvious that investing in housing is sub-optimal as it is quite logical for an investor to accept the lower return on housing if they do not consider the extra return from the stock market to be sufficient to compensate for the additional risk and effort. The effective annualised return of the assets provides another perspective of their relative merits as it shows the effect of the differing variances of returns. The greater volatility of the sharemarket results in its effective return falling the most from its average rate of return. Interestingly, the consistency of the housing returns leads to its return improving relative to the sharemarket’s return. Graphically, this is shown in the following cumulative returns graph. Figure 6: Cumulative Real Rates of Return on Different Assets for the Period, 1970-1997. Source: Westpac FPG 8 Shows the compound annual real rate of return for an investment made in 1970 and held until 1998 whereas the other column shows the arithmetic average of annual returns. Real Cumulative Returns 0 100 200 300 400 500 600 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1970 set as 10 0 real nzse40 real 6 month dep rate real 10 yr g ovt bond Adjusted housin g index 16 Phase 3 of the project will standardise the returns from the different assets by appropriately correcting for their risk premiums. Figures 5 and 6 as well as table 3 are for New Zealand as a whole. The regional experiences of housing returns may differ from one another. It is proposed to investigate the regional housing experience in Phase 3. A further feature of a housing versus shares comparison is that individuals typically hold completely undiversified housing portfolios, that is they have one house in one region. It is possible to estimate risk and return for the share and housing indexes but the figures for the share index are likely to be a far better proxy of the figures facing the average investor than those from the housing index as investors can easily hold a diversified share portfolio (eg by investing in an index fund). Therefore, despite the fact that the share index may be more volatile than the housing index it cannot be assumed from this that the typical individual’s housing holdings are less risky than their share holdings. An implication of this is that a risk adjusted return for housing based on these indices will be overstated for the average individual. ii) Cumulative Returns for Different Time Periods The 28 year period chosen for the graphs above relates to the period for which we have been able to obtain data. This period, however, does not necessarily coincide with the returns data which most people would consider relevant when attempting to forecast future returns. In this case, the data from the last 5 to 10 years could be expected to be more influential to peoples’ investment decisions. This is important as the starting point matters when making cumulative return comparisons. This can be seen in figure 7 using the last ten years of data. Figure 7: Cumulative Real Rates of Return on Different Assets for the Period, 1987-1997. Source: Westpac FPG Real Cumulative Returns ( 1987 - 1997 ) 0 20 40 60 80 100 120 140 160 180 200 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 Index (1987=100) real nzse40 real 6 month dep rate real 10 yr govt bond Adjusted housing index 17 Over this ten year period the sharemarket was outperformed by all of the other three assets, primarily due to the significant losses it made in the 1989-90 period. Notice that the graph starts at 1987, so would already incorporate the stock market crash that occurred in that year. 18 5. THE IMPACT OF TAXATION ON INVESTMENT DECISIONS Table 2 above displayed the data available on New Zealand households’ investment portfolios. One of the variables likely to affect New Zealanders’ investment decisions is tax. If taxation alters the relative returns from the various asset classes it can be expected that taxation also alters the investment decisions of individuals, thus leading to investment patterns that are privately optimal but socially sub-optimal. Data on the marginal returns expected from the various asset classes is either non-existent or unsuitable. Therefore, the analysis of preand post-tax returns that follows is more illustrative than factual. In some places we will use historical data, while illustrative numbers will be used otherwise. Table 4: Effect of Taxation on Asset Returns Asset Nominal Pre-tax return (%) Nominal Posttax return (%) Housing 13.96912.9610 Funds held in M3 institutions 9.511 6.37 Private share holdings 27.412 18.34 Passive Investment Fund 27.4 18.34 Managed Fund 27.4 14.2713 The return to housing and interest on M3 institution funds are the average over the last 28 years. The last three assets in Table 4 represent fully diversified portfolios of New Zealand shares differing only in the manner in which they are held. In a no-tax world we have assumed that the rate of return is equal to the average rate of return displayed by the New Zealand stock market index over the last 28 years grossed up to reflect company tax already paid before it is reflected in the share price or dividends. Therefore, the figure in the pre-tax column for diversified share holdings is 27.4%, representing the after tax average return of 18.34% grossed up for the 33% tax rate. The working behind the share portfolio returns is as follows. The profit or earnings performance of the companies in the stock market index is such that if 9 Average annual nominal return (1970 – 1998) for residential New Zealand housing including an estimate of 2.5% per annum for implicit rental minus maintenance and capital improvements. 10 Although capital gains and imputed rents on owner-occupied housing are not taxed, local government rates are still payable (assumed to be 1%) and no deductions for interest expense or maintenance are allowed. This return figure could be lower for rental housing as rental income would be taxable, but deductions would be allowed for interest expenses, maintenance and depreciation. 11 Average six month deposit rate (1970 – 1998). Interest income taxable at 33%. 12 Average return for the NZSE40 (and previously Barclays Index) for 1970-1998, including dividend returns, and grossed up to reflect the effect of tax on company earnings. 13 Fund manager may be able to delay payment of tax until assets are sold. 19 they retained all that income, untaxed, within the firm, the stock price would increase 27.4%. They then pay 33% tax reducing the return to 18.34%. The firms decide to pay 6% (say) of this to investors as dividends and retain the remaining portion, thus generating an appreciation in the share price of 12.34%. The private investor and the passive fund are unaffected by any further taxation as the dividend they receive is fully imputed and the share price appreciation is not taxed. The managed fund, however, does have to pay tax at 33% on the share price appreciation, thus generating its lower return of 14.27%. The returns to housing figure drops by 1% to reflect local government rates but any capital gains and imputed rent are not taxed. Also, no tax deductions for interest expenses or maintenance costs are allowed for owner-occupied housing. These reduce the attractiveness of housing investment, particularly if mortgage financing is used. In the pre-tax world, individuals would be indifferent between investments in passive or managed funds, depending on their beliefs regarding the ability of a fund manager to beat the market. If they had sufficient capital to form a diversified portfolio privately, they would also be indifferent between belonging to a fund or holding shares privately. The weights of assets that investors held between housing, fixed interest and share investments would depend on the relative riskiness of those assets and the individual’s risk preference profile. Following tax being imposed, a number of changes occur in an individual’s investment behaviour compared to the pre-tax world. Shares are now less desirable compared to housing as the relative returns have been altered by tax. Within share investments, managed funds have been the most adversely affected. For an investor to receive an equivalent return between passive and managed funds, the return on a managed fund now has to be up to 50% more than that of a passive fund. Fixed interest is the asset class next hardest hit while owner-occupied residential housing is least affected by tax considerations so becomes relatively more desirable in a post-tax world. To a certain extent the significance of these statistics depends on the characteristics of investors in New Zealand. For example, many people may be prepared to invest in shares but do not wish to do so on their own account as they either do not possess the skills to do so successfully or they do not have sufficient capital to purchase a diversified portfolio. In this case, the lower posttax return for managed funds could make them opt for housing instead of a managed fund. The tax advantage, or disadvantage, of other assets over housing is not always clear cut. It depends on the circumstances of the particular comparison being made. We illustrate this complexity by comparing the tax status of an investment in passive funds with the tax status of investment in housing. 20 Certain “passive funds” have a tax advantage over other financial investments in that the distributions of their capital gains are not taxed14. The tax treatment of their capital gains is similar to the tax free status of capital gains on housing. These passive funds can claim deductions for expenses earned in the production of taxable income. These deductions provide the funds with a partial tax advantage over home ownership because the home owner cannot claim tax deductions for interest and other expenses. Some features of the tax system can favour passive funds while other features favour housing. For example, passive funds are disadvantaged compared with housing because the passive fund must pay tax on their other income less costs and the tax therefore reduces their disposable income. The home owner does not pay tax on the surplus of imputed rentals less ownership costs. The disposable income from that surplus is not reduced by tax. Consequently, this feature, of itself, provides home ownership with a partial tax advantage over investment in passive funds, even though neither pays tax on capital gains and only the passive fund can claim tax deductions for costs. By contrast, the following feature of the tax system may reverse this, placing home ownership at a partial tax disadvantage when compared with commercial investment. Business losses from one period can be offset against taxable income in future periods, lowering later tax liabilities. Sole traders can offset business losses against other income. Home owners cannot use any deficit in imputed rents less interest and other house costs either as an offset against other income or against future taxable income. Early interest costs under a table mortgage are high and could exceed imputed rents, involving imputed losses on the housing asset. Home ownership is at a partial tax disadvantage compared with commercial investment as a result of this. A house owner experiencing imputed losses early in the life of a table mortgage is offsetting them against any capital gains and is waiting for the long term benefits that arise when the interest payments under the table mortgage is small. In general, the net effect of the tax system is that when housing has a surplus of imputed net rents home ownership has a tax preference compared with a passive fund incurring a trading profit. However, when both are in loss15 the passive fund has the tax benefit of losses carried forward. The size of the cost and revenue items involved for the house and the fund determine where the tax advantage lies in each comparison of the tax status of different assets. 14 These are certain Group Investment Funds with category B income. 15 A housing investment may well be in a loss early in a table mortgage. 21 a) Leveraging Investments The leverage in a house purchase can be defined as the ratio of the mortgage to the downpayment. This leverage ratio has subtle effects through table mortgages, where principal repayments are initially low but rise through time, and vice versa for interest payments. The interest commitments of a mortgaged owner-occupier are met from other income. This is one of the self-limiting features of leverage, an individual cannot increase the leverage on a house without limit. The amount that can be borrowed is constrained by market income. The greater is the leverage ratio, the greater are the interest payments involved in the purchase of a given house. These interest payments are a burden on the homeowner. Repayments of mortgage principal reduce this burden but also involve a cost because the amounts repaid are not available to buy assets that would yield the market rate of return. Lower mortgage interest payments are to be set off against market income foregone as capital payments are made. The difference between the mortgage rate and the financial funds rate of interest is therefore important. New Zealanders typically leverage themselves quite aggressively to purchase real estate. As an example, it is now possible to obtain 95% financing from some banks. For comparison, note that the minimum deposit typically required in Japan to purchase residential housing is 30%, thus allowing only 70% financing.16 On the other hand, it appears to be relatively rare for New Zealanders to leverage themselves for the purpose of investing in the share market. In addition, credit institutions are prepared to lend more aggressively against housing than against shares due to the differing volatility of the prices of the assets and consequent security of the collateral. The main reason people borrow to purchase housing is that housing is an expensive item that they do not have sufficient capital to afford. They are required to borrow in order to spread payment over a long period. While this may be the main reason for leveraging the investment, the level of borrowing does have an effect on the risk and return on the housing purchase. In phase three of this project we shall estimate a risk-adjusted rate of return for a leveraged housing investment. 16 Poterba 1994 22 6. FURTHER RESEARCH Stage three of this project is to involve more detailed research into areas of interest or hypotheses identified during stage two. Further work has been identified as follows: • Locate data on assets and wealth that have been omitted from the available statistics on household wealth holdings to ensure that comparisons can be accurately undertaken. • Locate saving and household net wealth data on countries which have similar GDP per capita levels to New Zealand in order to be able to make more meaningful saving comparisons. • Investigate methods for improving the returns to housing data by better incorporating rent and imputed rent into the returns series rather than the approximation that we have used, and investigate regional housing price indices. • Risk-adjust the rates of return from the four main asset classes used in this paper. • Undertake more sophisticated analysis of international saving rates after adjusting for income levels, drawing on OECD work • Test the hypothesis that investment behaviour with respect to housing is similar across countries and confirm the factors driving this investment behaviour. • Undertake a more sophisticated analysis of the effective tax rates on different forms of investment to assess how the tax system is affecting New Zealand’s investment behaviour. 23 7. CONCLUSIONS This project summarises the current state of our knowledge on saving behaviour in New Zealand and makes some comparisons with the saving statistics available for other countries. Historical rates of return for the main asset classes in New Zealand appear to indicate that housing has been a strong performer in comparison to stocks and fixed interest investments. While average rates of return are higher for the stock market, if imputed rents and other rental income from housing are included, housing has compared favourably over the last thirty year period. Additional factors such as taxes and leveraging behaviour may be emphasising this trend. Internationally, the statistics we have been able to gather appear to indicate that New Zealand annual household saving as a proportion of household disposable income is low by comparison with other OECD countries, but annual national saving is broadly comparable. Net worth as a percentage of disposable income is also within the range of international comparison but generally on the low side. One concern is that the available statistics make some serious omissions which could be significantly affecting the international comparisons. Stage three of the project will include attempting to improve the statistics on household assets, particularly including items such as rental housing, commercial property, and equity in non-listed businesses including agribusiness. Statistics indicate that the perception that New Zealand has a high proportion of home ownership compared to other OECD countries is true. The interesting feature of the housing statistics, however, is that differences in home ownership ratios across most OECD countries are generally small but the proportion of total net assets represented by housing assets is far greater in New Zealand than elsewhere. This result is sensitive to the possible statistical omissions identified above. If the statistics are an accurate representation, one possible hypothesis is that individuals in the countries represented are actually following a similar investment rule but that other factors such as the differing cost of housing as a proportion of disposable income result in differing investment holdings. Further research to test this hypothesis would be valuable and would require finding data on average incomes, house prices, and other consumption goods prices in the range of countries to be compared. Changes in social factors such as average household size may also be having an effect. Statistics indicate that average household size in New Zealand has been declining significantly leading to an increase in the amount of housing per person consumed.