Outgoing RBA Governor, Phil Lowe, hopes the Australian economy can remain “on an even keel” but stating that “the path to achieving a soft landing remains a narrow one” which sounds like we’ve already run aground. So, should you bet on a recession?

There’s nothing ‘technical’ about a ‘technical recession’

A frequently used (but arbitrary) shorthand is that two consecutive quarters of falls in GDP is a ‘technical recession’. But this simplistic definition is often misleading, as discussed below.

In the US the impressively titled National Bureau of Economic Research’s ‘Business Cycle Dating Committee’ dates when recessions occurred. They determine the timing of recessions based on range of monthly data (income, employment, consumption, retail sales and industrial production). Two of the eight members joined when it was formed in 1978, so they’ve certainly seen a few business cycles.

Dating recessions in Australia

Even without a panel of esteemed academics we can still do better than the crude rule of two negative quarters of growth. There are a couple of sensible alternative definitions.

• The first is to identify peaks and troughs in the business cycle using an algorithm applied to the level of per capita GDP (per person better captures people’s standard of living). This shows there were three recessions in the 2000s. Australia’s claim to 29 years without a recession doesn’t stack up. This method also shows the early 1990s recession lasted 9 quarters, much longer than the ‘technical’ recession definition. Anyone who lived through that recession knows it was long and deep
  
  
• A second alternative is to date recessions with the unemployment rate. We care about the effects of recessions on people, and job losses are the most tangible impact. A widely used option is the ‘Sahm rule’, which says a recession is when the three-month average of the unemployment rate has increased by half a percentage point in a year.

Both alternative definitions of recessions have advantages. The business cycle definition produces much more intuitive dating and lengths for recessions, and it is still based on aggregate output (i.e. GDP). However the unemployment definition is simpler and it is much more timely. There is always a long lag in dating a recession with GDP. We need two quarters of GDP and GDP is published with a long lag. Recession can even be over before we can say for certain that it started! By contrast employment data is published monthly and with less lag. So, the unemployment definition will tell us we’re in a recession sooner, even if employment growth tends to lag GDP growth.

Recessions across countries

Countries often experience recessions at the same time. This can be because of common shocks, say the 1970s oil price shock or the 2008 financial crisis. Also, downturns in large economies spill over to others through trade and investment flows or confidence effects. For many years it’s been said that when the US sneezes Australia catches a cold. Since the 1980s, every recession in Australia has occurred when half or more of other countries are also in recession.

The probability of a US recession

The slope of the yield curve (the 10-year interest rate minus the 3-month interest rate) has been shown to be a good predicter of US recessions. Recessions often follow a big increase in the Fed Funds rate, while prospects of future lower growth and interest rate cuts result in lower long term interest rates. All US recessions have been preceded by short-term interest rates that are high relative to long-term rates.

Leading indicators – combinations of forward looking economic and financial series (such as orders for goods, building permits and equity prices and bond yields) – are another useful predictor of recessions.

Both the yield spread and a high-profile leading indicator are at levels only previously seen in recessions. A statistical model based on the yield spread suggests the probability of a US recession in one year is 66%. Adding in the leading indicator pushes that up to 81%. More commentators might be saying that the likelihood of a US recession is declining given the resilience of the economy, but if you back the models, you’d still put your money on a recession.

The probability of an Australian recession

We can estimate the likelihood of a recession in Australia using the slope of the yield curve (10-year minus 3-month or overnight interest rates) and the OECD leading indicator (which comprises construction approvals, manufacturing surveys, share price index, terms of trade and 10-year bond yield). High inflation or very low unemployment can also suggest a recession is brewing.

Our statistical model indicates that the probability of a recession in Australia in one year is 43% (using the unemployment definition). Using the ‘business cycle’ definition it’s even higher at 60%. But whether Australia does have a recession is likely to depend on whether the US does. If the US does experience a recession, then there’s a two-thirds chance Australia has a recession, but if the US sails through then the odds of an Australian recession fall to one-in-three.

Not all recessions are equal

While the odds of an Australian recession might be one-in-two, not all recessions are equal. The early 1980s recession saw the unemployment rate rise 4 percentage points while it ‘only’ increased by around 1½ percentage points in the 2008 GFC recession.

Recessions are deeper when there are amplification mechanisms. Often this comes from banks’ earlier bad lending decisions. This was an important factor in the severe early 1990s recession. In recent years Australian banks have maintained good lending standards, and like insurers they are well capitalised and have plenty of liquid assets.

One area that has been highlighted as a potential risk in Australia is the high level of household debt. We’re a world leader in that race. At almost 190% of income, our household debt is more than two-and-a-half times as large as in the 1990s deep recession. But from what we know, those households that owe debt are mostly well placed to meet their payment obligations. It’s likely that highly indebted households will significantly cut consumption in a recession, but I don’t expect that will be enough to turn a mild recession into something much deeper.

Overall, I think there is a one-in-two chance that Australia will have a recession, but my best guess is that it won’t be severe.