Clear evidence of US recession

I haven't commented before now on the US labour market stats which came out a week ago---I've been kept busy with changing my data sources and rewriting my programs to work with these new formats.  Plus other software improvements, as well.

First, non-agricultural employment.  Note how employment growth started to pick up in the second half of last year, and has been falling since the beginning of this year.  Employment growth is still positive, but only just.  Ignoring the Covid Crash, it hasn't been this weak since coming out of the GFC, 15 years ago.

Second, unemployment.  This comes from a different survey to the payrolls data.  The BLS gets the payrolls data by asking companies how many people they employ.   They get estimates for unemployment by asking a random sample of households whether they are employed or unemployed.

If these two different surveys, drawn from different samples populations, show the same thing, we can be more confident about what's happening.   And they do.

The unemployment rate is usually regarded as an indicator which lags the cycle, i.e., it turns up or down after the economy does.   However, its change over 6 months coincides quite well with the cycle, except it is inversely correlated---it rises when the economy falls.  So in the chart below, I have plotted the six-month change inverted.  A falling line with this indicator thus indicates a slow-down or a recession.

Observe how unemployment had started falling (shown as a rising line in the chart) in the second half of last year, showing that an economic recovery was getting underway, and how this recovery has reversed since January.

Another data series from the household survey is total employment.  Because this comes from the survey of households, not employers, it shows a slightly different picture to the payrolls chart.  But not that different.  And it also points to a very rapid slowdown in jobs since January---the largest fall since the GFC, if we exclude the Covid Crash.  Note that because the household survey draws on a smaller proportion relative to its sample population than the payrolls survey, its random month-to-month fluctuations are larger.   So I have used a six-month average change to smooth this.

Like the ISM surveys, all three charts show that a recovery in the economy had begun, and this recovery was aborted by Trump's tariff imbroglio.

I see no reason for this to change direction over the next 6 months, because we will now be seeing the inflation effect of the huge jump in tariffs.  This will reduce real incomes, and therefore expenditures.

Will the Fed cut rates to save the day?  No.  Not until it's sure that the rise in inflation from tariffs is transitory.  And even if it does, interest rate changes take many months to increase economic activity.

There will be random month-to-month zigs and zags, but I expect US economic data to worsen inexorably for the next few months.