Symposium on Inequality: Causes and Consequences June 19, 2014
Keywords: Distributional Economics; Statistics;
The purpose of this paper is to present some data on top market incomes in New Zealand. It updates an earlier paper by one year to 2011/2 and extends the estimates it back to 1936/7 .
The background to this paper is the international top incomes data base assembled by Facundo Alvaredo, Tony Atkinson Thomas Piketty and Emmanuel Saez. This paper provides an implicit critique of their New Zealand series without minimising their statistical achievement nor criticising Piketty’s theoretical analysis.
Working with data is tedious; definitions require great care. I shant talk in this paper much about methods, but I need to discuss definitions,. I do by differentiating between alternatives.
New Zealand. Some of the public discussion confuses what is happening in New Zealand with what is happening elsewhere (especially the US). Too often we imitate what is going on overseas instead of applying general principles to the New Zealand specifics. Towards the end of the paper, I place the New Zealand results in an international context as far as possible.
Natural persons who are adults (over 15). As far as possible the data to be presented excludes trusts and companies and other such legal artefacts.
Covering all adults deals with where the data base does not include all income recipients. While taxpayers were 98.2 percent of the adult population in 2012, in 1936/7 only 12.0 percent of adults where taxpayers.
This standardisation also allows us for the impact of women joining the paid labour force – a very important post-war phenomenon. I’ll say something more about adults with odd residential status.
Wealth or Income?
Income. We dont not have very detailed information on top wealth.
Disposable or Market Incomes?
Ideally we would like to report market incomes. In practice the data being used is income reported for income tax purposes. Basically, it is market income but there is a little contamination from National Superannuation. Some omissions are explained below.
To be Compared With?
Not all market income is reported for tax purposes particularly in the past when not all individuals filed tax returns, and Inland Revenue had no other means of identifying their income. Instead we have used private market incomes as measured in the National Accounts. Unfortunately there is only a detailed series back to 1980/1. I have projected back earlier using National Accounting estimates of private income.
How Far Back?
To 1936/7. Earlier tax data does not separate out those who are not natural persons, such as companies.
Consistent Through Time?
‘Fraid not. I’ve done my best. What I particularly cant do is allow for changes in tax law.
One tax change which complicates the data is the treatment of corporate dividends. Until 1989 they were ‘doubled taxed’. Corporations paid tax on their profits and their dividends paid from the tax-paid profits were treated as taxable income of the shareholder. From 1989 there has been a dividend imputation system in which a shareholder receiving a dividend from a company is entitled to an ‘imputation credit’, which represents tax paid by the company and is offset against the shareholder’s income tax liability. In effect corporation tax becomes a withholding tax for shareholders’ dividends.
This altered the way that dividends are recorded by the IRD. For example, $100 of corporate profits which were taxed at, say, 33 percent and fully paid out were recorded as $67 before imputation but as $100 after the new regime was introduced. Thus the taxpayer’s recorded income went up, but so did their after tax income (by the same amount).
In order to get consistency over time I have treated the grossing up of these dividends as the substantial tax break that it was, so we are classifying the increase as belonging to an increase in disposable income . (That was in addition to the substantial lowering of the top tax rate from 1988.)
What Part of the Income Distribution?
Top incomes only. I shall report the income shares of the top 10 percent, 1 percent and 0.1 percent of adults. Also the Pareto coefficient which I explain shortly.
Given these Limitations, Why Bother?
Because it is there, I suppose. The reason it is prioritised is that Piketty’s book considers what is happening to top market incomes. What has been happening in New Zealand?
I am reporting is the best data I have. Robert Solow famously justified some statistical work he was doing by citing the addicted gambler who knew ‘the casino wheel is crooked but it is the only one in town’. At least he knew what he was doing.
The following 2012 tax year benchmarks may be useful:
There were about 3.5 million adults over the age of 15. So the top 10 percent of income recipients amounted to 350,000, the top 1 percent were 35,000 and the top 0.1 percent were 3,500.
10 percent of adults had an income above about $72,500 and a 37.4 percent share of all income.
1 percent of adults had an income above about $165,000 and a 9.7 percent share of all income
0.1 percent of adults had an income above about $500,000 and a 2.7 percent share of all income..
The annualised average wage was around $45,000 while the average adult income was $36.000.
The Pareto Coefficient
Vilfredo Pareto famously proposed that upper incomes followed a power probability law characterised by a single parameter, the ‘Pareto coefficient’. They indicate how compressed the top tail of a distribution is. The lower the coefficient the more unequal is it is – the more stretched out
Pareto coefficients are in excess of 1, but typically near 2. If the coefficient is 2 and there are 1000 above income $X, then there will be 250 above $2X. But if the coefficient is 3 there will be only 125 above that income, the smaller number indicating the distribution is more equal.
The strength of the Pareto coefficient is that it represents well the top of a distribution, while the rest of the distribution need not be known. Its weakness is the converse.
The figure shows the Pareto coefficient for top incomes between 1936/7 and 2011/2. Initially it starts low at around 2.0. The average of 17 OECD countries in 2005 came to 2.1; on this measure New Zealand was about as unequal at the top before the Second World War as is typical for an OECD country today.
The coefficient then steadily rises to about 3 by 1960. Over the entire period it averages about 2.9; high compared to many other countries, which implies a lower top inequality. It then runs at this three-ish level from the early 1960s to the end of the 1980s, after which it perhaps begins to rise.
I explain the reasons for this pattern after I have looked at income shares.
Top Income Shares
The pattern for the top 10 percent of adults is the converse of the story of the Pareto distribution. They have a high share of around 35 percent of all market income from just before the War. A 35 percent share means that the decile had an average income 3.5 times the national adult average. A 25 percent share is 2.5 times the national average.
The 35 percent level continues to about 1959/60, and then falls to 25 percent in about 1980. The top decile’s share then stagnates through the 1990s since when it has been increasing slightly.
The patterns for the top 1 percent and top 0.1 percent are broadly the same as for the top decile, except that there is no evidence of their share increasing in the last two decades.
Observe that there is little evidence in the data of a a business cycle – perhaps surprisingly. There is probably not a lot to be gained from a year-to-year analysis because of sampling variability. However some changes need to be dealt with, before we look at the trends.
The 2000 Blip
There is a definite blip in the income shares in the 1999/2000 year indicating an increase in inequality. The top income tax rate was increased from 33 to 39 percent for the 2000/2001 year. Many taxpayers arranged their income flows to move income from the high tax year back to the lower tax year, temporarily raising income in 1999/2000 and lowering it in the following year.
Why Did Inequality Measure at the Top Decline in the First Part of the Post-War Era?
There was a secular decline of the share of top Incomes in the first 40 years after the Second World War.
Its causes were probably more related to the remaining 90 percent of adults and cannot be tracked from this data basis. However other work I have done suggests that the most important driver was the impact of full employment in the period. It operated through the following four channels.
– Male labour force participation rose, essentially out of unemployment.
– Female (paid) labour force participation rose dramatically as changing household circumstances and domestic technologies made it easier to (also) work outside the home.
– Maori migration from the countryside into the urban centres, which increased their market incomes.
– There seems to have been compression in remuneration margins within the labour force.
What Happened After 1990?
Since the 1980s the early post-war drivers towards less inequality were no longer there. Full employment, as we understood it in the early post-war era, no longer exists. Probably our ‘normal’ level of unemployment will be now similar to other rich market economies. The post-war migration of woman into the paid labour force and Maori into cities is largely over. The institutional mechanisms which enabled wage compression have been largely abandoned. I shall have more to say about this. So the increasing inequality which characteris4ed the first four decades of the Post-War era came to an end. The market (tax assessed) income inequality largely stabilised.
However the share of the top 10 percent seems to have marginally increased (although it is volatile) suggesting some increases in overall inequality. Yet the Pareto coefficient also incrementally increases, which is in the opposite direction suggesting a reduction in inequality among top incomes.
The two results can be reconciled if the strong increase in shares is accruing to those in the second to tenth percentiles – the top 10 percent less the top 1 percent (The 2 to 10 top percentiles.) . That would compress the top of the income distribution as indicated by the mildly rising Pareto coefficient.
Since the very top is far more influenced by rewards to capital, while below them there is a greater impact from the remuneration to top managers and professionals it would seem that in the last few decades the rewards to ordinary labour. Piketty observes this effect too.
One local factor may have been the 1988 State Sector Act which abandoned the rigid relativities that existed in the public service, enabling higher relative remuneration to the top civil servants, while most civil servants were experiencing restricted real increases (or declines). The same thing was happening in the private sector; a consequence of the globalisation of the market for management and higher professionals.
Unfortunately we cannot estimate the magnitude of the margin increasing in order to assess to what degree that explains the rest of the upshift.
If this hypothesis is correct then the driver of the recent increasing inequality is from widening labour earnings rather than increases in the return and quantity of wealth.
What About the Piketty Thesis?
First notice is that New Zealand’s high income recipients have low incomes compared to those overseas. Our top 0.1 percent are about 3,500 individuals who report annual taxable incomes of $500,000 or more in 2011/2. Around 700 would report incomes in excess of $1,000,000.
New Zealand does not seem to follow the Piketty thesis of rising inequality in top incomes. But this would be to adopt the Piketty thesis crudely.
New Zealand has no sophisticated financial sector. That means no mega-remunerations. (There is no general agreement within the economics profession as to why this is happening.)
But the Piketty effect is even more explained by patterns of wealth accumulation and returns. We dont have the New Zealand data to explore this directly. Before tackling the issue I need to explore some other possibilities.
There are, of course, measurement problems. The data series since 1981/2 are of higher quality. However there are omissions.
The data series does not cover trusts. Apparently trusts have become more common since the ending of inheritance tax in 1992.
Nor does it cover capital gains. There is often confusion about the effect of omitting capital gains. Unquestionably including them would increase the level of income inequality. On the basis of the handful of countries for which there are estimates, the inclusion of capital gains might add about 1 percentage point to the share of the top 1 percent’s income – say increasing their share from 5 to 6 percent of private income.
However, while the omission of capital gains reduces the measured inequality, it does not automatically follow that it disguises increasing inequality. It is not impossible that capital gains were smaller after the Global Financial Crisis than before it. In which case inequality of top incomes may hardly been have changing at all. We just dont know.
My suspicion is the big issue which makes the data difficult to interpret though is what may be called ‘partial New Zealand residents’.
Partial New Zealand Residents
Under New Zealand tax laws, those with high incomes can avoid declaring offshore income for taxable purposes by avoiding being New Zealand tax residents. The criteria for being a New Zealand tax resident are
– living in New Zealand for more than 183 days in any 12-month period, or
– having an ‘enduring relationship’ with New Zealand, or
– being away from New Zealand in the service of the New Zealand government.’
People who are not New Zealand tax residents are liable for New Zealand tax only on their New Zealand-sourced income. Thus they may appear in the IRD data base but only part of their income is reported.
New Zealand is such a small economy that those with very large fortunes are likely to hold wealth portfolios diversified by jurisdiction. It is not implausible that as little as a third of their income comes from New Zealand sources; only that part is reported in the tax statistics.
Given increasing international mobility it seems likely that an increasing proportion of those at the very top of the income distribution are not tax residents. If so, any Piketty effect of a growing elite of the rich is likely to be missed in the New Zealand tax data.
Earlier I argued for a New Zealand social science, and not an imitative colonial one. But New Zealand analysis needs always to be in the context of a globalised world.
Politics and Market Incomes
Disraeli summarise privilege as ‘pay, patronage and power’. Recently there have been increasing public concern about the extent that those on top incomes are influencing the political process.
Underlying this concern is the ideal of democracy being about ‘one person one vote’, whereas market activity is about ‘one dollar one vote’. In practice the two areas of public life cannot be so easily separated, so one can infringe excessively upon the other. For instance, it it is now generally accepted that before the mid-1980s, politics was too involved in market decisions. But can the opposite happen? This is an evident political concern in the US; does it apply in New Zealand? This is a wider issue than this paper can cover; here are few pointers.
It is an interesting feature of New Zealand’s electoral system that we now have three minor parties openly backed by millionaires. Each is dependent upon the threshold effect which our MMP system allows. Many think it is an anomaly; perhaps it becomes even more anomalous if it enables millionaires to buy seats in parliament.
Perhaps political donations are more in the spirit of democracy if they are transparent. It is not obvious they are sufficiently transparent in New Zealand.
The rich have also the ability to buy acolytes to promote their political views. Again transparency of funding sources may be vital, but as one who is unwilling to curtail open speech it seems to me that it would be better to develop institutions with an alternative view rather than have the lopsided funding of lobbying which currently dominates New Zealand.
It also appears that some of those who are not tax residents play a significant role in New Zealand political life as donors, as political advocates and as lobbyists (and as voters). Given that taxation is the price of citizenship is this appropriate? Perhaps such political activities amount to having an enduring relationship with New Zealand.
The share of those with top incomes fell up to the end of the 1980s, while top incomes became increasingly compressed. Shortly after, there were increases in inequality arising from increases in remuneration margins for management and professionals and the introduction of a dividend imputation system. There have been small or no increases in inequality since.
Calibration difficulties make international comparisons difficult, so we must be cautious about ranking New Zealand’s top income inequality with economies elsewhere.
However there is no evidence of a major surge in inequality in the New Zealand data in the first decade of the twenty first century, as has occurred in the UK and the US, probably because New Zealand does not have as sophisticated financial sectors as they have and because New Zealand’s wealthy may function – for some purposes – outside the country.