top of page

Using unemployment rate as an indicator of recession

Edited By: Lavanya Goswami


This piece analyses the “Sahm Rule” proposed by former economist at the US Federal Reserve, Claudia Sahm. Specifically, the usage of movement in unemployment rates as an indicator of recession, and the merits of this approach in policy-making to safeguard against recessions. This article also looks at the flip side of the argument–how this approach could create excessive pessimism around recessions and thus justify potentially harmful policy choices. The article concludes by looking at alternative symptoms of recession–thus challenging the idea of the Sahm Rule being a ‘law’ rather than a historical pattern.

Introduction: What is the Sahm Rule?

The Sahm Rule is a policy instrument for disbursing automatic stimuli to households when the economy is approaching a recession. As per the rule, the US government should automatically make direct payments to households as soon as the three-month moving average (that of the current month and the two preceding ones) of the national unemployment rate is above 0.5 percentage points of the lowest point of the preceding twelve months.

Essentially, if the unemployment rate appears to be rising above the past year’s trend, that should act as a trigger for automatic stabilisers, which help households smooth over variations in individuals’ income through the recession period, and thus props up consumer demand. This boost to demand is a useful counterbalance to recessionary tendencies–which are inherently those of weak demand.

Why unemployment as an indicator?

As Sahm herself points out, unemployment rate is a useful indicator of “labour market strength and overall economic well-being.” This intuitively makes sense: a trend of growing unemployment implies an ongoing economic slowdown, and more importantly, upcoming economic slowdown. On the one hand, the onset of economic slowdown could lead to falling output growth, falling incomes, falling demand, falling prices, and ultimately more people out of work as economic activity becomes less profitable. On the other hand, rising unemployment due to exogenous factors (war, disease, financial crises etc.) leads to falling consumer demand, which sets off the same cycle as above, leading to falling output and ultimately even more unemployment.

So why not just use output (simply, GDP), as an indicator of recession instead? Sahm points out the important fact that GDP indicators are far more lagged, and therefore lack the timely quality that make unemployment a relatively more useful indicator. After all, if the government is to release direct stimuli early on in the recession, it cannot retrospectively predict recessions based on months-old GDP data.

Further, Sahm provides robust historical data analysis backing her very specific trigger–the three-month moving average of national unemployment being 0.5 percentage points higher than the lowest point of the past twelve months. She looks at all US recessions in the 1970-2018 period–seven in total–to calculate this trigger. She observes: “even a modest rise in the unemployment rate such as 0.50 percentage points…has occurred only during or closely following recessions.” Thus, the trigger for automatic stimuli would kick in early on–generally three months–in each recession since 1970.

The downsides of the ‘rule’

In her more recent writings, Sahm becomes more cautious about employing her ‘rule’ for predicting recessions. Most significantly, in the context of the post-COVID experience of the US economy, she warns us about the dangers of treating her ideas as a “law of nature” rather than an “empirical regularity.”

Following the massive fiscal deficits employed by the US government to fight the lockdown-induced recession during the pandemic, the country has faced massive inflation. As we observed through the past year, the US Federal Reserve has followed an aggressive contractionary policy to fight inflation–it has raised interest rates eleven times since March 2022, so far, with more hikes expected. This has of course triggered fears of a ‘hard landing’–a fresh recession deliberately induced by the central bank (in this case, the Fed) in order to control inflation.

Recall that while inflation is characterised by ‘excess demand’, a recession is characterised by ‘deficit demand’. Thus it stands to reason that a recession would break the back of persistent inflation (and the other way round). In this case, the severity of the post-COVID inflation has [ostensibly] merited discussions about a ‘necessary’ hard landing.

However, is a hard landing necessary? Many economists in the US would say yes, but the problem lies in how the debate is framed–specifically when the ‘Sahm Rule’ is invoked. Policymakers such as Larry Summers–former Treasury Secretary of the US–have invoked the Sahm Rule to justify the necessity of a hard landing. According to them, any Fed policy to fight inflation will lead to some degree of increase in unemployment–which is mostly true. Rising interest rates are designed to slow down economic activity, which logically leads to rising unemployment. At this point, they invoke the Sahm Rule to ‘predict’ that recession will inevitably follow this rising unemployment. In short: the Sahm Rule would imply that anti-inflationary policies, or attempts at ‘soft landings’ will inevitably lead to ‘hard landings’–recessions with high unemployment. Since a hard landing is ‘inevitable’ there is nothing we can do about it. Or so the argument goes.

Of course if policymakers adopt such an attitude, then a recession is inevitable, since nobody wants to explore the alternative realities where a ‘soft landing’ is in fact possible–less inflation but without a recession. And Sahm also rightly points out the moral hazards at play–none of the economists advocating for hard landings will be facing job losses, but ordinary working citizens. Importantly, Sahm points out that her ‘rule’ is not an unchanging law of nature–if in fact there can ever be such a thing in economics–and is purely an empirical finding with historical precedent. In the post-pandemic paradigm, policymakers would do well to explore the possibility of this precedent breaking.

Alternative explanations

Why, then, may unemployment not be the best indicator of recession? Or at least, not an infallible indicator? For one, recession may not be inevitable following rising unemployment. Particularly in the post-pandemic era, other factors, such as high household savings, are likely to prevent the onset of a vicious recessionary cycle. The way Sahm has been emphasising over the past year on ‘breaking the Sahm rule’, other economists such as Udit Misra–a columnist for the Indian Express–also talk about the possibility of an economy growing in “a way…that there is rising unemployment.”

Secondly, and more relevant to the Indian context, is the potentially misleading nature of the unemployment rate. While it is traditionally a useful measure of evaluating the labour market of an economy, it becomes less so when labour force participation itself becomes worryingly low. Essentially, the unemployment rate only measures the percentage of people looking for work but do not get it. It does not account for those who do not seek employment in the first place. India’s labour force participation rate (LFPR) in the 2022-23 financial year dropped to 39.5%. Therefore, even a falling unemployment rate is not necessarily cause for celebration–it may just be because of fewer people looking for work than more jobs being created.

The overarching conclusion from this discussion is of course that in economics, there can be no hard and fast laws of nature. Rules and stylised facts are often, if anything more than a theoretical reality, no more than historical precedents. And we know better than to expect history to always repeat itself.

Recent Posts

See All

Girl Math 101: Believing that it's real

Written By: Kuhuo Bajaj | Edited By: Lavanya Goswami I am a well informed, financially aware, modern woman. Of course I do girl math. Purchases that are returnable for a refund aren’t just free, but p


bottom of page