Published daily by the Lowy Institute

Is the relationship between growth and inflation shifting?

The RBA leadership is clearly thinking about the possibility that the relationship between inflation and output growth may have altered in an enduring way.

Photo: Daniel Munoz/Getty Images
Photo: Daniel Munoz/Getty Images

With all of the focus on interest rates, sometimes fundamental assumptions underpinning monetary policy are overlooked in the commentary. At times like this, when there are tentative but unmistakeable signs of possible change in those fundamentals, it’s worth stepping back to look at the big picture.

As widely expected the Bank of Canada increased its policy rate last month but otherwise the global trend to higher policy interest rates lost momentum. Neither the Reserve Bank of Australia nor the Bank of England are likely to raise rates at meetings this week, the US Federal Reserve deferred another rate rise at its meeting last week, and European Central Bank officials have hinted that any monetary tightening is a long way away. In all these cases, central banks observe that growth and employment are improving in their economies, but inflation remains below the rates they were accustomed to before the 2008 financial crisis.

In Australia, recent speeches by Reserve Bank of Australia governor Phil Lowe and deputy governor Guy Debelle directly addressed this issue in an Australian and global context. The talks have been widely interpreted as affirming that the RBA is also in no hurry to increase the policy rate, and anyway won’t move in ‘lockstep’ with other advanced economy central banks. That interpretation is certainly right, but it is also incomplete. In their speeches both Lowe and Debelle also raised wider and much more interesting questions about what the central bank is targeting, and why.

Though both officials reiterated the Bank’s commitment to the 2% to 3% inflation target the speeches directly or by implication raise questions about whether that target may one day need to be adjusted, and how the trajectory of policy rates may be influenced by a changing relationship between output growth and inflation.

Because they cast their comments in terms of global forces, the speeches are interesting not only in terms of Australian monetary policy, but also for other advanced economies and their central banks. As Debelle pointed out, other advanced economies mostly share with Australia the puzzle of disappointing output growth as well as low wages growth and low inflation compared to formal or informal targets.

Most advanced economies now face the possibility that low inflation may be indefinitely prolonged, even as output growth and employment pick up. It is most evident in the US, where inflation remains well below the Federal Reserve’s informal target, wages growth is sluggish, yet unemployment is very low. At around 2%, output growth is not very far below a sustainable long term rate.

An inflation target was selected in Australia in the early 1990s for various theoretical and practical reasons but in the years since the rationale has changed. An inflation target these days is seen as a reliable barometer of whether or not economic output and employment are growing at a sustainable rate. If inflation is rising beyond the target, growth is probably too fast to be sustainable. If inflation has fallen under the target, then output growth is probably too slow.

This rationale is rarely spelt out, but it is evident from the post meeting statements, published minutes of board meetings, and from the speeches of RBA officials, that the Bank is mostly concerned with the rate of growth of output and employment. Inflation is usually treated as a consequence of the rate of growth of output and employment rather than an objective that can or should be controlled independently of output growth. With different shades of emphasis this is true of most advanced economy central banks. There are islands of dissent but most central banks act on the belief that they can affect interest rates, that interest rates affect output and employment, which in turn affect inflation.

It follows that, if there is one, the inflation target selected should be one that experience discovers is compatible with a sustainable rate of long term output growth. For the last 20 years in Australia the target of inflation between 2% and 3% has proved to be amazingly consistent with average output growth a shade above 3% and average consumer price inflation of 2.5%.

The rationale behind inflation targeting

At the core of inflation targeting is a belief that wages growth would accelerate as unemployment fell, and as wages growth rose so would consumer price inflation. Higher interest rates could slow demand, drive up unemployment, lower wages growth, and ultimately lower inflation. More broadly, the underlying idea is that if output is increasing beyond its sustainable rate, inflation will pick up. If there is surplus capacity, inflation will slow. Higher interest rates slow the economy and thus slow inflation. This is by no means the only channel of influence for monetary policy. In the Australian case, monetary policy can change the exchange rate and directly affect inflation. But the connexion between inflation and the rate of growth of output compared to its sustainable rate is important.

Both the Debelle and Lowe speeches tentatively raise the possibility that the relationship between inflation and output growth may have changed in an enduring way. If this is so, it is obviously a very big issue for monetary policy.

Debelle’s speech discussed the reasons Australia’s GDP growth rate and inflation are now subdued. These include slower workforce growth and slower productivity growth compared to 20 years ago. They also include trends that may or may not be transitory and are also apparent in other advanced economies. These are modest growth in business investment, the lingering impact of the 2008 financial crisis on attitudes to risk, and (for inflation) slow wages growth. All up the result in Australia and elsewhere is both lower trend output growth and lower inflation. 

Debelle does not know whether lower inflation in Australia and elsewhere is merely temporary or long lasting. Nor at this point does anyone else. Federal Reserve chair Janet Yellen, like Lowe and Debelle, stick to forecasts that inflation will sooner or later return to the trend rate evident before the 2008 financial crisis.   

Is slow wage growth here to stay?

In his speech on Wednesday last week Lowe took these monetary policy issues a little further. He pointed out that jobs growth in many advanced economies ‘has generally surprised on the upside’ so that in those countries ‘the unemployment rate is at, or below, the rate conventionally associated with full employment. Yet at the same time, growth in wages remains subdued, even in countries with low unemployment rates.’ He wondered ‘what does this mean for the outlook for inflation and monetary policy?’ Though it is often thought to be transitory Lowe pointed out that slow wage growth might turn out to be ‘much more persistent’.

Lowe then asked whether the relationship between wages growth and unemployment was changing in Australia, as it appears to be changing in other advanced economies. 

If wages growth is less sensitive to employment growth and this change was enduring, it should imply I think that the steady rate of inflation associated with a sustainable rate of output growth will be lower than we have come to expect. It should also imply that if the Bank is to pursue an inflation target it should be lower. It is not possible to tell in Australia right now because low wages growth is new phenomenon. 

Lowe posed the most interesting and consequential question to arise from these two well considered talks. If wages growth is lower for any particular rate of unemployment (technically, if the Phillips curve is flatter) then ‘how hard’ Lowe asked, should the Bank should press to ‘get inflation up’? 

If wages growth is a proxy for inflation and unemployment is a proxy for output growth, then Lowe is also pointing to the possibility that a sustainable rate of output growth may now, and in the future, be associated with a lower rate of inflation. It is possible therefore that the current inflation target will prove to be too high. 

[[{"fid":"184701","view_mode":"default","fields":{"format":"default","alignment":"","field_file_image_alt_text[und][0][value]":false,"field_file_image_title_text[und][0][value]":false},"type":"media","field_deltas":{"2":{"format":"default","alignment":"","field_file_image_alt_text[und][0][value]":false,"field_file_image_title_text[und][0][value]":false}},"link_text":null,"attributes":{"height":463,"width":800,"class":"media-element file-default","data-delta":"2"}}]]

Photo: Carla Gottgens/Bloomberg via Getty Images

More broadly, more radically, more portending for the current inflation target, the relationship between output growth and inflation may have changed. Let us say for the moment that potential output growth is now 3%, composed of about equal contributions from the growth of hours worked and the growth of output per hour worked. If unemployment in Australia falls without stimulating more rapid wages growth, the RBA might find the economy well on the way to regaining 3% output growth, without reaching 2.5% consumer price inflation. 

A quarter century ago there was a dramatic and persistent decline in Australian inflation which was as unexpected as it was welcome. There is no particular reason to suppose that a much less dramatic but equally persistent decline in the rate of inflation consistent with maximum sustainable output growth might be possible. Indeed, it may have already occurred. The changing mix of output and demand in advanced economies, the decline of trade unionism, the abundance of commodity resources compared to the growth of demand, the combination of modern technologies with low labour costs in the emerging economies, the shock of the 2008 financial crisis – any and all of these influences might contribute to a prolonged period of low consumer price inflation in the advanced economies, even if household and business demand strengthens. 

In practice the objective of monetary policy is to maximise output growth consistent with inflation remaining within the target band on average. The RBA has never faced the problem that output growth might be at potential, yet inflation remains below the target band or at its lower bound. 

This then brings us back to Debelle’s talk the previous Friday. The neutral policy interest rate is the one that balances savings and investment at full employment. In this space a month ago we suggested that if the RBA forecasts of 3% output growth and 2.5% inflation by the end of 2019 were attained, then a companion ‘neutral’ policy rate would be 3.5% - 2 percentage points higher than the current policy rate of 1.5%, but at least one percentage point lower than the neutral rate over the last couple of decades. In the minutes of its July meeting, it turned out that a far more rigorous investigation by RBA staff had come up with the same number. 

Possible impact on the ‘neutral’ policy rate 

But in the meeting minutes and again in Debelle’s talk the Bank has been careful to cast the ‘neutral rate’ in real or after-infation terms. The Bank suggests the real rate is 1%. If the inflation target was 2.5% and attained, then the neutral nominal rate – the one we observe – would be 3.5%. But specifying a real rate instead of a nominal rate leaves it open for the Bank to gradually accommodate itself to a discovery, not yet openly broached let alone confirmed, that potential output could be sustained with a rate of inflation below the current target.  It could perhaps be where both underlying and headline inflation are now, just under 2%. In that case the ‘neutral’ policy rate would be 3%, not 3.5%.  It would also mean we are closer to what the Bank may come to define as an acceptable inflation rate than we thought. 

Together the speeches affirmed that potential output growth in Australia was lower now than it was twenty years ago. It is at least possible that a steady rate of inflation associated with that rate of growth of potential output might also be lower. If it is, then the implication is that either the inflation rate specified in the target will need to change, or perhaps entirely reformulated. 

It is not an issue the RBA needs to bother too much about right now, but I think we need to be alert to it. If a change of emphasis is occurring then the focus should be on output growth rather than inflation. The reason the first rate rise is still quite a way away is not that inflation is 1.8%. That in time might be thought a quite acceptable target. The main reason is that output growth in the year to March was 1.7%, the lowest four quarter rate since the global downturn in 2009. That is the number to watch. If output comes up to 3%, with unemployment no higher than now, it seems to me the tightening episode will probably have begun even if inflation is still 2%. 

Lowe and Debelle are right to insist as they vehemently do that the current inflation target remains in place. It is after all an agreement between the Bank and the Treasurer, and it is much too early to say that Australia will return to an optimal rate of output growth but fail to reach the current inflation target. There will be plenty of time to debate the alternative targets and policy rules. The Bank meanwhile has a good deal of practical flexibility, as Lowe emphasised in his remarks. After all, if the current inflation target was interpreted literally, the policy rate today would not be 1.5%. It would be zero, or close to it.  

Inflation targeting is not entirely satisfactory, but for Australia and for now it is better than the alternatives. While the maximum sustainable rate of growth of output is always the underlying objective of monetary policy, it is quite difficult to use as a target. The problems involved in  defining potential growth – particularly in terms of defining the rate of growth of productivity – are why the RBA typically refers to ‘trend growth’, with the usually unstated inference that the past average is close to potential. 

The RBA could over time move to an explicit target of sustaining output growth at potential so long as inflation did not persistently run at over 2.5%, or perhaps 2% if that turns out to be the new average outcome. That would eliminate the need to bother about low inflation if output growth was OK. But it would make uncomfortably explicit the RBA’s responsibility for real output growth and employment. Since it can influence only interest rates and the exchange rate it would be reluctant to accept explicit responsibility for sustaining output growth. It would also mean the RBA would need to do a lot more work on calculating and predicting changes in potential GDP. 

Another possibility is nominal GDP targeting, advocated by ANU and Brookings economist Warwick McKibben. Its major problem for Australia is that volatile export prices mess with nominal GDP. In the year to March nominal GDP in Australia increased by a very lively 7.7%, but underlying consumer price inflation was under 2% and output growth not much more than half of potential. Two years earlier, four quarter nominal GDP had increased 1.4%. Both numbers were misleading signals about what really matters, which is the output of goods and services. 

For the future trajectory of monetary policy and interest rates much depends on what now happens to wages and inflation, compared to real GDP – and not only in Australia, but also in the United States, Europe and the UK. But in these two recent speeches, the leadership of the RBA is clearly thinking about the possibility that the relationship between inflation and output growth may have altered in an enduring way, with important consequences for monetary policy.

You may also be interested in