Who ate all the pie?

A week after Australia’s federal election, the Fairfax newspapers featured an article, “Numbers say we’re growing quite nicely“, by the economics editor of the Sydney Morning Herald, Ross Gittins. As an overview of the last few decades, this was a work of casual lustration:

The mildness of last year’s recession means the economy has now entered its 20th year of growth since the deep recession of the early 1990s. But how has this growth been distributed through the economy? [...] If you imagine it’s got a lot more unequal, then you’ve been reading too many newspapers.

Gittins has been at this before, criticising the popular memory of John Howard as “the prime minister who did so much to widen the gap between rich and poor”:

But just remember: this widely held view – firmly believed by many ordinary voters as well as the ideologues of the left – exists independent of the evidence. When you trouble to examine that evidence you find the story far less damning than people imagine.

The journalist got one thing right: you shouldn’t believe what you read in the newspaper. For his argument was based entirely on coarse measures of inequality, like the Gini coefficient, derived from periodic household surveys. Due to sampling constraints, these surveys don’t capture information from the apex of the income pyramid: a tiny minority of households. They give a cut-off qualifying point for the top decile (the upper 10 percent of the income distribution, those individuals earning over $75 000 in 2007), but tell us nothing about how wealth is shared within this well-off layer.  As more income is concentrated in the upper tail of the distribution, therefore, the less useful such surveys become as a measure of inequality.

When we look at Australian taxation statistics, on the other hand, the last few decades evince – just as people imagine – a dramatic reassertion of the power and privileges of the ultra-rich. The time series below is from data reconstructed by Tony Atkinson and Andrew Leigh. It displays the share of total personal income accruing to the top 0.05 percent of the income distribution (these days some 5000 people out of 10 million taxpayers) in each year since 1941. As we can see, after declining steadily throughout the postwar decades, the income share of the super rich reversed trajectory around 1980, leaping from 0.6% in that year to around 2.5% in 2007.

A similar though less pronounced restoration occurred across the entire top decile of the income hierarchy (percentiles 90 to 100), where the falling shares of total income during the 1960s and 1970s were steadily recovered over the last three decades. But within this shared pattern, these four fractiles diverge somewhat. The higher reaches (P99-100) rose quickly and have surpassed their earlier shares. But the income share of P95-99 both declined and increased less spectacularly; that of P90-95 stayed comparatively constant.

When we add the great mass of the population (the lower 90% of the personal-income distribution) to this picture, it becomes clearer still. The series below (normalised to 1965) shows the respective changes in the share of income received by various fractiles since that year. Note the inverted U-shaped curve of P0-90, which increased its share of the income pie until about 1980, after which it declined drastically. This is mirrored almost exactly by a compensating movement in the share of the top 1%, which followed an opposite U-shaped path over this period. Note, however, that P0-90 is made up of about 9 million people, while P99-100 includes roughly 100 000 members. Income gains by the top 1% over the past three decades have been spread across a much smaller group, and therefore carry greater weight for each individual.  

When, as happens sometimes, this trend of rising inequality is publicly acknowledged by economists or journalists, it is often ascribed (as our old pal Gittins did here) to ”skill-biased technical change“. Innovation has favoured the “knowledge worker” and frowned upon the manual worker; their respective levels of material reward follow from this. But, as shown above, in recent decades the incomes share of fractiles P90-95 has risen barely at all. Income inequality is due not to the rising salaries of software engineers, but the engrossment of the richest within the rich. Unless we propose that top CEOs and coupon clippers possess some ineffable Trump-like skillset that the lower 99% of the population lacks, the income-follows-marginal-productivity argument is impossible to sustain.

Instead, one explanation for the divergent fates of the fractiles shown above lies in their respective income composition. P90-99 constitutes the upper-salaried layers of Australian society: most of their income (though less than for P0-90) arrives in the form of salaries or wages. The rich of the rich, however, depend more heavily on capital income (dividends, interest, rent, royalties), capital gains and partnership income.

Thus, during the 1980s, the growing income share of the uppermost percentile (P99-100) followed naturally from:

1. skyrocketing real interest rates (from zero in 1977 to about 10% in 1989)
2. a recovery in the average profit rate
3. a reduction in retained profits and corresponding increase in profits paid out as dividends
4. the subsequent increase in the price of financial assets (the influence of sharemarket spikes in the late 1980s and 1990s can be seen clearly in the income shares above).

To this list we may add cuts by the Hawke-Keating government to the top tax rate, and changes to the institutional form of the state, which was moulded to the requirements of this ultra-rich rentier layer. Price stability became the core statutory objective of the Reserve Bank; while the the Prices and Income Accord enforced wage “restraint” and a decline in unit labour costs (exchanged for growth in the assets of union-controlled superannuation funds).

Alongside the boost to capital income, one of the notable trends behind the rising income share of the top fractiles is a rise in the importance of wages and salaries. The graph below shows that, alongside the year-to-year asset-price fluctuations that raise or lower non-salary income, salary income has risen steadily. The latter now accounts for a large portion of the income-share received by P99-100. The most publicised aspect of this trend has been the tremendous growth of CEO salary packages (usually including stock options) relative to average wages. The function of these high salaries is – obviously enough – not to reward toil or managerial acumen; but still less is it about securing the “best talent” from competitors. It is merely the most efficient way for (1) top managers to secure a share of the surplus product, given that a large proportion of profits are distributed as dividends; and (2) owners to ensure that managers share their strategic objectives. The ”wages and salaries” of top fractiles are thus, like capital income, merely another form of earnings linked to the ownership and control of assets. Where top managers do not already hold such assets, the outsized scale of executive remuneration soon transforms them into owners. The growth of salary income in the share of the top percentile thus suggests a merger at the top of Australia’s social hierarchy, allowing the ready circulation of personnel between ownership, company boards and top management.

What, then, accounts for the stability of a social order in which a tiny minority (a few tens of thousands of individuals) has in recent decades secured such a large slice – the greatest since the Second World War – of total personal income?

For one thing, the allegiance of upper-salaried layers (P90-99) has been secured by a rise in income relative to the broad mass of the population (P90-99). For another, returns on financial investments now contribute substantially to the retirement income of this stratum, and no doubt serve an ancillary ideological purpose, linking the interests of the working rich to those of high finance.

Tags: Australia, executive pay, finance, inequality, Marxism, media, neoliberalism