Speech The Glass Half Full

It is very good to be back in Adelaide. Thank you for the invitation.

As we meet here, economic discussion in Australia has reached a rather curious position. Consider the background. Australia avoided a deep downturn in 2009, when most countries did not. A large number of businesses and jobs were saved by that outcome – though we will never know how many.

Almost as a matter of arithmetic, the ensuing upswing was always going to be of the moderate variety. Rapid cyclical growth usually comes after a serious slump (and when it doesn't, it comes just before one). After small downturns, less spectacular growth is the usual experience. So it has proved on this occasion.

Even so, three and a half years after the depths of the crisis in late 2008, this unspectacular growth has nonetheless seen real GDP per capita well and truly pass its previous peak. This is something yet to be achieved in any of the other nations shown here (Graph 1).

Graph 1
Graph 1: Real GDP per Capita

According to data published this week by the Australian Statistician, real GDP rose by over 4 per cent over the past year. This outcome includes the recovery from the effects of flooding a year ago, so the underlying pace of growth is probably not quite that fast, but it is quite respectable – something close to trend. Unemployment is about 5 per cent. Core inflation is a bit above 2 per cent. The financial system is sound. Our government is one among only a small number rated AAA, with manageable debt. We have received a truly enormous boost in national income courtesy of the high terms of trade. This, in turn, has engendered one of the biggest resource investment upswings in our history, which will see business capital spending rise by another 2 percentage points of GDP over 2012/13, to reach a 50-year high.

To be sure, we face considerable structural adjustment issues arising from the mining expansion, and from other changes in the world economy. These are not easy to deal with (though they are not insurmountable). And we live in a global environment of major uncertainty, largely because of the problems of the euro zone. Nonetheless, an objective observer coming from outside would, I think it must be said, feel that Australia's glass is at least half full.

Yet the nature of public discussion is unrelentingly gloomy, and this has intensified over the past six months. Even before the recent turn of events in Europe and their effects on global markets, we were grimly determined to see our glass as half empty. Numerous foreign visitors to the Reserve Bank have remarked on the surprising extent of this pessimism. Each time I travel abroad I am struck by the difference between the perceptions held by foreigners about Australia and what I read in the newspapers at home.

I harbour no illusion that this can suddenly be lifted by anything I say today. But it is, hopefully, worthwhile to offer a few facts, and some perspective and analysis of the situation.

The Multi-speed Economy

Much of our public discussion proceeds under the rubric of the so-called ‘two-speed economy’. It's become very much the description of the moment, and not only in Australia. One picks up the same theme in many other countries. Indeed it is a description of the global economy. Growth in the advanced industrial countries continues to be sluggish, and in some cases output is going backwards. Within Europe, Germany has been doing well, while other nations face huge economic challenges. Meanwhile growth in the ‘emerging world’ has been pretty robust apart from the effects of natural disasters. So in popular terms, we might say that there are varying lanes on the global growth highway: fast, slow, very slow. There are a few economies in the breakdown lane.

Turning to Australia, we have long had a multi-speed economy. For example, it has been a very long-running trend that population growth tends to be faster in Western Australia and Queensland than in Tasmania or South Australia. Typically, certain industries such as housing construction show the expected differences due to population growth.

Moreover while we debate the rise of mining and the much heralded ‘decline of manufacturing’, we might note that it has been a very long running trend that output and employment in manufacturing has grown more slowly than in the economy as a whole, and that output of various kinds of service provision has grown faster. That has been happening for at least five decades, and in most countries in the developed world. In the case of Australia's manufacturing sector, this decline reverses an earlier rise. In fact, the respective shares of mining and manufacturing in Australia's GDP at present are about where they were in 1900.

It is obvious at present that the mining expansion is quite concentrated both in its industrial and geographical dimensions, and economic indicators do reflect that. But the mining sector is not the only sector growing. If the recent data are taken at face value, the non-mining economy has grown at about 2 per cent over the past year. Mining employment is indeed growing quickly – interestingly enough according to the available data, the increase in mining employment exceeded the fall in manufacturing employment over the past year. But the largest increase of all was in the sector called ‘health care and social assistance’, in which employment rose by about the size of the combined fall in manufacturing and retailing employment over the same period. And while there are clearly differing drivers by industry and by region, there are mechanisms that even out at least some of these differences. Spillovers do occur both in the private sector and via the tax and expenditure system.[1] Remarkably, in the face of the understandable concern about job losses in particular regions and industries, the dispersion of unemployment rates by statistical region is no larger today than has usually been the case over the past 20 years. Hence, while there are clearly multiple speeds, the total speed seems to have been one of reasonable growth and low unemployment.

The Behaviour of Households

But there is another aspect of the ‘multi-speed’ experience, which I suspect explains a good deal of the dissatisfaction we see, and it has to do with the behaviour of the household sector. Some parts of the economy that depend on household spending are still experiencing relatively weak conditions, compared with what they have been used to. But this isn't because the mining boom spillovers have failed to arrive. It is, instead, the result of other changes that actually have nothing to do with the mining boom per se, but a lot to do with events that occurred largely before the mining boom really began.

The story is summed up in the two charts shown below. The first shows household consumption spending and income, both measured in per capita terms, and adjusting for inflation (Graph 2).[2] In brief, household spending grew faster than income for a lengthy period up to about 2005. The arithmetically equivalent statement is that the rate of saving from current income declined, by about 5 percentage points over that period.

Graph 2
Graph 2: Real Household Income and Consumption

It was no coincidence that households felt they were getting wealthier. Gross assets held by households more than doubled between 1995 and 2007. The value of real assets – principally dwellings – rose by more than 6 per cent per annum in real, per capita terms over the period (Graph 3).

Graph 3
Graph 3: Real Household Assets

Only a small part of this was explained by an increase in per capita expenditure on dwellings. The bulk of it came from rising prices. Moreover, a good deal of borrowing was done to hold these assets and household leverage increased. The ratio of aggregate household debt to gross assets rose, peaking at about 20 per cent. There was definitely a large rise in measured net worth, but relative to aggregate annual income, gross debt rose from 70 per cent in 1995, to about 150 per cent in 2007. Correspondingly, by 2007 the share of current income devoted to servicing that debt had risen from 7 per cent to 12 per cent, despite interest rates in 2007 being below those in 1995.

It is still not generally appreciated how striking these trends were. I cannot say that it is unprecedented for spending to grow consistently faster than income, because it had already been doing that for the 20 years prior to 1995. That is, the saving rate had been on a long-term downward trend since the mid 1970s. But it is very unusual in history for people to save as little from current income as they were doing by the mid 2000s. And it is very unusual, historically, for real assets per person to rise at 6 per cent or more per annum. It is also very unusual for households actually to withdraw equity from their houses, to use for other purposes, but for a few years in the mid 2000s that seemed to have been occurring.

Of course, Australia was not alone in seeing trends like this. There were qualitatively similar trends in several other countries, particularly English-speaking countries that experienced financial innovation. The international backdrop to this period was the so-called ‘great moderation’, in which there was a decline in macroeconomic variability. There were still business cycles but downturns were much less severe than in the 1970s or 1980s, inflation was low and not very variable, which meant that nominal interest rates also were generally low and not very variable, and compensation for risk became very modest.[3]

This ‘moderation’ came to an end with the crisis beginning in 2007. And with a few years of perspective, it is increasingly clear that Australian households began to change their behaviour at that time, or even a little before. The rate of saving from current income stopped falling probably around 2003 or 2004, and began to increase (we now know), slowly at first as the income gains from the first phase of the resources boom started in about 2005 or 2006, and then more quickly in 2008 and 2009.

Real consumption spending per head initially remained pretty strong in this period, reaching a peak in 2008. It then declined for a year or so, before resuming growth in the second half of 2009. That growth has, however, been much slower than had been observed previously. In the nearly three years from mid 2009 through to the March quarter 2012, real consumption per head rose at an annual pace of about 1½ per cent.

This is more than a full percentage point lower than the growth rate from 1995 to 2005. But this sort of growth is, in fact, quite comparable with the kind of growth seen in the couple of decades leading up to 1995. It is in line with the quite respectable growth in income. But the gap between the current level of consumption and where it would have been had the previous trend continued is quite significant. If we then consider the growth of foreign online sales and so on, and the fact that consumers seem more inclined to consume services – experiences, as opposed to goods – we can see this is a significant change for the retail sector.

No doubt reinforcing this trend towards more circumspect, but more typical, behaviour is that the earlier strong upward trend in real assets per head has abated over recent years. In fact, real household assets per head today are about the same as they were five years ago, with a dip during the crisis, a subsequent partial recovery and then a slow drift down over the past couple of years. Both dwelling prices and share prices – the two really big components of wealth – have followed that pattern.

At some point, wealth will begin to increase again. After all, people are saving a reasonable amount from current income and placing the proceeds into various assets (especially, of late, deposits in financial institutions). That is, they are building wealth the old-fashioned way. Ultimately these flows will be reflected in higher holdings of real and financial assets, at least once debt levels are regarded as comfortable. Asset valuation changes can, of course, dominate saving flows in shifting wealth over short periods and they are inherently unpredictable. So no one can predict the course of these measures of wealth over any particular short period. But wealth will surely resume an upward track, sooner or later.[4]

When it does, however, it is unlikely to be at 6 or 7 per cent per year in real, per capita terms. I would guess that over the long term, something more like 3 per cent would be nearer the mark.

I think this is a profoundly important point and worth emphasising. The decade or more up to about 2007 was unusual. It would be quite surprising, really, if the same trends – persistent strong increases in asset values, very strong growth in per capita consumption, increasing leverage, little or no saving from current income – were to re-emerge any time soon. That is, the gap between consumption today and the old trend level on the chart is not going to close. I noted to another audience about three years ago that the prominence of household demand in driving growth in the 1990s and 2000s was unlikely to be repeated.[5] If there were business strategies that assumed a resumption of the earlier trend, they will surely be disappointed in time, if they have not been already.

There were several parts of the economy that benefited from that earlier period, and that are finding the going much tougher now. Retailing was obviously one, but so was banking. Banks and other financial institutions enjoyed rapid balance sheet and profit expansion as they lent to households and some businesses. But they can see that period has now finished. Businesses that serviced rapid turnover in the dwelling stock (such as real estate agents, mortgage brokers) are seeing those revenue streams considerably reduced, and are having to adjust their strategies and capacity to suit changed conditions. For example, the rate of dwelling turnover is about one-third less than it was on average over the previous decade, and about half its peak levels. This is affecting state government stamp duty collections as well as the real estate sector.

We can also see some echoes of these changing trends in household demand in business investment spending.

Graph 4
Graph 4: Real Business Investment

This chart shows business investment, split into mining and non-mining, and measured in real, per capita terms, so as to be consistent with the earlier charts (Graph 4). Investment has been on a stronger upward trend since the mid 1990s than it had been for a number of years before that. In particular, business investment in real per capita terms has grown, on average, by over 6 per cent per annum since 1995, more than double the average pace over the preceding 35 years. Moreover a lot of this was in the non-mining sector, and it began before the present run up in mining investment really got going. Some of this growth reflected the same ‘consumer facing’ growth sectors mentioned above. Of the four sectors that had the fastest growing investment spending over that period, three were finance, one called ‘rental hiring and real estate services’, and retail trade. Some of these sectors are slowing their investment rates now.

Meanwhile, mining investment has recently been rising at an extraordinary pace. In 2005, mining investment was near its long-run average of around 2 per cent of GDP. By mid 2014 we expect it to reach at least 9 per cent of GDP. If that occurs, mining investment will be about as large as business investment in the rest of the private economy combined. As a result of that, total business investment will reach new highs this year, and next. Hence, there is a very large build-up in the nation's capital stock occurring. If it is well managed and soundly based, that ought to allow the possibility of further growth in output and incomes. The investment phase of the mining boom will start to tail off in a couple of years' time, after which the shipments of natural resources should step up significantly.

We might expect by then as well that some other areas of investment spending that are weak at present will be picking up. More generally, I suspect we will discuss the nature of investment quite a bit in coming years as we grapple with structural change in the economy and powerful shifts in the population's needs (think of investment in the aged-care sector, for example, or public infrastructure needs). We will also be looking for productivity pay-offs from the various investments.

But the key message for today is that the multi-speed economy is not just about the mining sector squeezing other sectors by drawing away labour and capital and pushing up the exchange rate. It is doing that, but slower growth in sectors that had earlier done well from unusually strong gains in household spending would have been occurring anyway, even if the mining boom had never come along. It is these changes in behaviour by households, in asset markets and in credit demand, that I think lie behind much of the disquiet – dissatisfaction even – that so many seem to have been expressing. But this would, as I say, have occurred with or without the mining boom. In fact, without the mining boom and its spillovers, we would have been feeling the effects of those adjustments rather more acutely than we do now. The period of household gearing up could have ended in a much less benign way.

Implications for Policy

What are the implications of these trends for economic policy, and particularly monetary policy? Does it have a role in helping the adjustment?

One thing we should not do, in my judgement, is try to engineer a return to the boom. Many people say that we need more ‘confidence’ in the economy among both households and businesses. We do, but it has to be the right sort of confidence. The kind of confidence based on nothing more than expectations of ever-increasing housing prices, with the associated willingness to continue increasing leverage, on the assumption that this is a sure way to wealth, would not be the right kind. Unfortunately, we have been rather too prone to that misplaced optimism on occasion. You don't have to be a believer in bubbles to think that a return to sizeable price increases and higher household gearing from still reasonably high current levels would be a risky approach. It would surely be a false basis for confidence. The intended effect of recent policy actions is certainly not to pump up speculative demand for assets.[6] As it happens, our judgement is that the risk of reigniting a boom in borrowing and prices is not very high, and this was a key consideration in decisions to lower interest rates over the past eight months.

Hence, I do not think we should set monetary policy to foster a renewed gearing up by households. We can help, at the margin, the process of borrowers getting their balance sheets into better shape. To the extent that softer demand conditions have resulted from households or some businesses restraining spending in an effort to get debt down, and this leads to lower inflation, our inflation targeting framework tells us to ease monetary policy. That is what we have been doing. The reduction in interest rates over the past eight months or so – 125 basis points on the cash rate and something less than that, but still quite a significant fall, in the structure of intermediaries' lending rates – will speed up, at the margin, the process of deleveraging for those who need or want to undertake it.

In saying that, of course, we cannot neglect the interests of those who live off the return from their savings and who rightly expect us to preserve the real value of those savings. Popular discussion of interest rates routinely ignores this element, focusing almost exclusively on the minority of the population – just over one-third – who occupy a dwelling they have mortgaged. The central bank has to adopt a broader focus. And to repeat, it is not our intention either to engineer a return to a housing price boom, or to overturn the current prudent habits of households. All that said, returns available to savers in deposits (with a little shopping around) remain well ahead of inflation, and have very low risk.

So monetary policy has been cognisant of the changed habits of households and the process of balance sheet strengthening, and has been set accordingly. As such, it has been responding, to the extent it prudently can, to one element of the multi-speed economy – the one where it is most relevant.

What monetary policy cannot do is make the broader pressures for structural adjustment go away. Not only are the consumption boom and the household borrowing boom not coming back, but the industry and geographical shifts in the drivers of growth cannot be much affected by monetary policy. To a large extent, they reflect changes in the world economy, which monetary policy cannot influence. Even if, as a society, we wanted to resist the implications of those changes other tools would be needed.

In fact Australia does better to accommodate these changes, and to think about what other policies might make adjustment less difficult and quicker for those adversely affected. It is in this area, in fact, that we need more confidence: confidence in our capacity to respond to changed circumstances, to respond to new opportunities, and to produce goods and services which meet market demands. It is also to be hoped that some of the recent positive data outcomes will give pause to reflect that, actually, things have so far turned out not too badly.


We face a remarkable period in history. The centre of gravity of the world economy seems to be shifting eastwards – towards us – perhaps even faster than some of the optimists had expected. Granted, that is partly because the relative importance of Europe seems to be shrinking, perceptibly, under the weight of its internal problems. But even if the Europeans manage the immediate problems well, there is no mistaking the long-run trend.

That this comes just as a very unusual period for household behaviour in Western advanced countries (including Australia) has ended, has been a remarkably fortuitous combination for Australia. Certainly it means we have the challenge of adjusting our behaviour and our expectations to new drivers for growth and new imperatives for responsiveness, but we do so with growing incomes, low unemployment and exposure to Asia. That is infinitely preferable to the sorts of adjustments that seem to be the lot of so many others at present.

The Australian community has understood that we can't base growth persistently on falling saving and rising debt and that is forcing changes to business models. But it has to be said that the return of a certain degree of thrift actually strengthens our medium-term position. If we can marry that to a focus on incrementally improving the way we do things – lifting productivity – there is actually a lot to look forward to. For Australians, the glass is well and truly half full.


See Lowe P (2012), ‘The Forces Shaping the Economy over 2012’, RBA Bulletin, March, pp 85–90. [1]

These are updated versions of charts I first used one year ago. See Stevens G (2011), ‘The Cautious Consumer’, RBA Bulletin, September, pp 77–82. [2]

There was, of course, a nagging problem of periodic financial panics. But several of these seemed to be managed without serious lasting damage. The Asian financial crisis was devastating for the Asian countries involved, but the global economy was not badly affected. The Russian crisis of 1998 – described, remarkably, by one experienced observer at the time as the worst since the 1940s – was similarly handled without serious fall-out. The bust of the dot-com bubble was associated with an economic downturn in the early 2000s but this too was, by historical standards, quite mild. Perhaps people began really to believe that major downturns were always avoidable and that higher leverage therefore was safe. If so, they had a major fright from 2007 onwards.  [3]

In considering these trends in wealth and household spending behaviour, we could ask which way causation ran – did changing wealth drive changing spending patterns, or was it the other way around? The answer is almost certainly that causation ran both ways. If rising asset values creates a sense of greater wealth and people feel less need to save from current income to achieve any goal they might have for their assets, they can spend more from current income. But in spending more, and being prepared to borrow more, they also tend in the process to affect asset values for both real and financial assets, which then reinforces the trend in wealth, and so on. So it is not possible, in a very simple analysis such as the one presented here, to disentangle all that. But it seems the two trends have been related, and mutually reinforcing, in both directions. Both the strong rise in spending and the strong rise in gross assets (and leverage) ended some years back now. [4]

Stevens G (2009), ‘Challenges for Economic Policy’, RBA Bulletin, August, pp 10–16. [5]

As in 2009, the challenge is ‘how to ensure that the ready availability and low cost of housing finance is translated into more dwellings, not just higher prices’. See Stevens G (2009) ‘Challenges for Economic Policy, RBA Bulletin, August pp 10–16. [6]