The intensity of my toddler’s demand for strawberries is a solid match for the fervent curiosity investors have about the current economic status of the US. While his fit of sulking over berries came to an end, questions and worries regarding the recession persist, with ambiguity in economic data stoking the uncertainty. This has been heightened by the appearance of various indicators pointing to the ‘now-cession’.
The Sahm rule, an indicator developed by economist Claudia Sahm for initiating fiscal stimulus, indicates potential tumult. Since 1970, any rise surpassing 0.5 percentage points in the unemployment rate over three months, as compared with its minimum in the prior year, has been associated with a recession. Alarmingly, this sign was triggered in July.
Investor tantrums are disruptive and loud. Due to this, analysts are keen to develop additional measures. UBS, for instance, has drawn on the employment-population ratio to reassure investors that a recession is yet to arrive.
In an effort to decipher the current economic condition, Pascal Michaillat from the University of California, Santa Cruz, and Emmanuel Saez, University of California, Berkeley, merged a mildly altered Sahm rule with a similar job change rate indicator, fashioning the Michez rule (a moniker I gave, not them). This rule has successfully detected downturns swiftly and it points to a recession starting way back in March.
It’s not promising when both Sahm and Michez rules warn that the recession has perhaps already set in. Even though the past is filled with unfounded claims stating ‘this time is different’ (I have learned not to be swayed by desperate cries professing my son’s newfound resistance to tart fruits), these indicators do indeed caution the current times.
Consequently, the conventional link between unemployment and job vacancies seems to have skewed in recent times, indicating that history might fall short in shedding light on the present situation.
Begin by comprehending the Sahm rule. As highlighted by Ernie Tedeschi from Yale University, the spike in unemployment seen from January to June 2024 was majorly due to individuals newly entering or rejoining the labour market. This suggests an increase in unemployment caused by a surplus in labour supply, rather than a perilous drop in demand. Additionally, in July we also observed a surge in short-term unemployment, which according to Tedeschi, warrants caution.
Caution must also be exercised with the Michez rule, which like the Sahm rule, leans on fluctuations in the unemployment rate. Although in the past, data on job vacancies was an effective measure of the labour market’s state, it has recently potentially given misleading signals. This could be as a result of a decrease in job availabilities that are merely indicative of normalisation following an unprecedented peak. In more recent years, the correlation between unemployment and job vacancies seems somewhat irregular, suggesting that past occurrences may not serve as a reliable guide to the current situation.
In essence, the situation at hand is rather uncertain. We could be experiencing a recession, although various other signs suggest otherwise. Is there a way to reflect this ambiguity quantitatively?
Michaillat and Saez have attempted this by setting two benchmarks, one at the bottom where certainty climbs from zero to 100 per cent, and one at the top. The lowest benchmark, crossed in March, is the lowest point their indicator has ever reached during all recessions since 1960. The upper benchmark represents the highest point their indicator has ever hit during all past recessions.
Based on these thresholds, the likelihood that we are currently in a recession is equal to the quotient of the distance covered between these two thresholds by the current indicator. According to the latest figures, this probability is 40 per cent.
Let’s consider an analogy. If my son has historically developed a stomach ache after consuming six strawberries, but the pain has always occurred when he has consumed a minimum of ten strawberries, this indicator would imply a 50 per cent probability that Caspar will experience a stomach ache after eight strawberries.
If this still seems perplexing, truth be told, I am in the same boat. Other than the question of whether the scholars are truly implying probability instead of some form of “confidence measure”, this approach appears to be highly sensitive to the temporal scope selected. For instance, if they had traced back to the 1930s, their upper benchmark would have been lower, endorsing a 67 per cent certainty of a current recession.
The core dilemma centres around the fact that since 1960, there have only been nine instances of economic downturns. This is insufficient to spot reliable trends that would unquestionably apply during the abnormal era following the pandemic. Therefore, unavoidably, my son and investors may feel some degree of irritation. All rights reserved for The Financial Times Limited 2024.