US leading indicators signal recession

For various reasons, I haven't calculated my US cyclical indices for a year or more.   But I've finally updated my US data banks and psyched myself up to do the calculations, so, here goes.

The chart below shows the year-on-year percentage change in my US coinciding and my US leading indices.  They are calculated from many underlying time series and are designed to remove some of the noise caused by the plethora of indicators which move in different directions each month, and that way to give clarity about the direction of the economy.

I have plotted my leading index with a 12-month lag.  This gives us an implicit forecast of the economy's direction over the next 12 months.  Observe how covid screwed up the lags, which is logical, because the covid crash and the recovery from covid were caused by exogenous influences, not by movements in the economy itself.

Note how the percentage change in my leading index is falling fast, suggesting that over the next 6-12 months the economy will be weak, or in recession.

The chart below compares my US coinciding index with my US diffusion index.  A diffusion index measures what percentage of a universe of monitored time series is rising.  In this case, the universe is 57 different time series, almost exclusively monthly.  When all are rising, the economy is strong.  When all are falling, it's in deep recession.  It's been smoothed using a 12-month centred moving average to iron out the monthly ebbs and flows.

It leads the cycle by about 5 months.  The unsmoothed diffusion index ticked up in November, but (a) that's based only on those data which were available, and (b) small blips in diffusion indices can be revised away as more data become available, and (c) it's just one month.  However, if this is the low for the diffusion index, it nevertheless indicates that, for at least the next 5 or 6 months, the US economy will be slowing.  

None of these indices gives pin-point timing or extent of the swings in the business cycle.  However, they do give strong rough indications of what's happening.  

My guess is that the US economy will be weak or even declining until the middle of next year.   But as I have said before, this is the first recession in my long professional experience caused by the extreme incompetence of the party and politicians in office, and by damaging policies, rather than by the strong ebbs and flows of the economy, so who knows?

US leading index falls

This is the month-on-month annualised percentage change in my leading index.  The data for the last couple of months are provisional.   Normally, I would regard this as not long enough to call a recession, but since the causes of the recession are not changes in monetary policy, but an artificially induced collapse in confidence as a result of Trump's tariff policies, which I don't see ending any time soon, I think it's more plausible to argue that the economic data are now confirming my a proiri expectations.

July labour stats in the USA

 I talked about the blip in the US PMI/ISM data here.

These charts show the labour force data.  The first one shows the change in payrolls, the second (from a different survey), the change in the unemployment rate (inverted, because unemployment goes up in recessions and down in recoveries)  In each chart, they are compared with the average of the manufacturing and services ISM surveys, extreme-adjusted.

You can clearly see the post-covid economic recovery in 2021, and then the gradual subsidence as fiscal stimulus wears off, high interest rates take effect, and the post-covid "revenge buying" in services slows.

Will the US go back into recession?  I don't think it will, because recessions occur when there is excess---excess debt, excess inflation, a property crisis, fierce response by central banks, a collapse in consumer and/or business confidence.  But it's relatively easy for slowdowns to happen.   In an uptrend, the economy can advance more or less rapidly, producing mini-cycles.  In recessions, mini-cycles are much rarer.  When the economy plunges, it doesn't pause to take its breath.  

The bottom chart shows the rate of change in my own US leading index, designed to give forward indication of changes in direction in the economy, compared with the year-on-year change in real GDP.   As you can see, the rise in the rate of change of the leading index should be consistent with a rise in GDP growth.  And it is.  Although it is showing a small downturn, indicating a slowing in GDP growth in 2025, it's not pointing towards an imminent recession.  

But my leading index doesn't include any measures of fiscal stimulus.  And that is tailing off:  if you look at the change in the Federal deficit, the IRA led to a big increase in the deficit in 2022, and a modest retracement in 2023.   That would have provided fiscal stimulus in 2022, and fiscal contraction in 2023.  Enough to cause a recession?  No.  But enough to take the shine off the growth numbers.

So, not a recession, but very likely a sluggish recovery, with a short-term blip.

My US leading index has turned up

 My US leading index, designed to give a year's warning of impending changes in the business cycle, has been rising since May.   The (lagged) correlation with the cycle has been excellent, up until the Covid Crash and its consequences.  And the gap between what my leading index predicted GDP should have been doing and what it actually did has widened.   Why?

My leading index is a combination of monetary measures, momentum measures of the economy, sentiment, and a variety of other leading indicators, all of which have in past led the business cycle.  What it doesn't contain is the impact of fiscal stimulus or contraction, and what it can't forecast is what economists call exogenous factors, such as Covid, war, weather, etc.   

The gap between what GDP should have done and what it did is due, I think, to two factors.  The first is the massive fiscal stimulus of Biden's "Inflation Reduction Act".  This didn't just cause plenty of government spending, which directly, and rapidly, expands overall economic activity, it also stimulated private sector spending, especially in the renewables sector.  The second is the burst of "revenge spending" as people took advantage of the end of lockdowns to travel, eat out and go to shows. 

In the economy, everybody's spending is someone else's income.  So stimuli have ripple effects, which only gradually tail off.   If a government builds a new freeway, those who work on it have incomes, the cement and steel companies have higher sales, and hire more people, and so on.  This higher income leads in turn to more spending, which leads in turn to more income.  You get the picture.  This means that these ripple effects can amount to much more than the original stimulus, particularly if business confidence is increased.  

The ripples of both these stimuli will fade.  For fiscal stimulus to keep on expanding the economy, it must be steadily increased, and that won't happen.  So growth will slip to more normal levels, but only slowly, as the "ripples" fade.  Offsetting that is the improvement in leading indicators, so growth will likely remain reasonable all this year.

In a non-recession year, incumbents tend to win:

First-term incumbency typically provides an advantage — unless there’s a recession during or just before the election. When there is no recession, the incumbent has always won in the post-World War II era. Goldman Sachs Research estimates a 15% probability of a recession over the next 12 months (equal to the average historical probability).

(Source: Goldman Sachs)

My US longer-leading index is troughing

First tentative signs that my longer-leading index for the US is bottoming.  Given the typical --- though variable! --- lag of 24 months, that suggests a low point in the US cycle of end 2024.  This may be brought forward if the Fed cuts rates aggressively, but it prolly won't because inflation is still a problem.  

Remember: the lags are long.  As we used to say in South Africa:  the economy doesn't turn on a tickey.  It takes time for monetary policy to take effect, in either direction.   The reaction of the economy to the sharp rise in the Fed Funds rate over the last year is already built in to the next year.   And a banking crisis will only make matters worse, although (good news!) that will discourage the Fed from raising rates any more.

But at least the index is bottoming.

My US leading index still declining

 My very own US "medium" leading index (which leads the cycle by about 1 year) continues to decline.  In the chart below, it is plotted as a deviation from trend, and with a one-year lag, giving an implicit forecast of economic activity over the next 12 months.   It shows no sign that it is bottoming.  

Shading shows recessions.

Expect GDP to fall from now on.

Clicking on the chart will give a much clearer image.  Goddess knows why; I do not.

My US leading index falls again

 My US "medium" leading index fell again in November.  It’s been plotted in the chart below as a percentage of its moving trend, and has been lagged (moved forward) by twelve months, giving us a good idea of the behaviour of the real economy over the next year: deepening recession.

Warning: the US recession has only just begun

 

My US leading indices point to recession

 My "medium"-leading US leading index now confirms my judgement that there will be a recession in 2023.

My US longer-leading index, which leads the cycle by 2 years, does not have as close a fit as the index shown above.  In the chart, it has been plotted with a 24-month lag.  It points towards a deep and prolonged recession.  Will that happen?  I've spent many years constructing these indices, and learning from forecasting mistakes, and I'm reasonably confident that the US will have a deep recession next year.  We shall see!

Meanwhile, ALL my leading indices point down.  My shorter-leading index, which leads by just a few months, suggests that short-term, the economic data are going to continue to deteriorate.

Deep recession -- if my US longer-leading index is right

This is my longer-leading US index, which leads the cycle by 24 months.  It suggests a deep 2024 recession, but the lag between the longer-leading index and the overall cycle might be shorter this time because of war and the commodity price explosion.  The lags between this leading index and the economy are quite variable, running from 18 to 30 months.  I expect the US recession to start in the next couple of months, but, as ever, it's entirely possible I am wrong.

At any rate, the longer-leading index is pointing towards the deepest recession in 35 years.  We shall see how accurate it is.

My tweaked "medium" US leading index

In this report, I commented that definitional changes to US money supply data caused a huge jump in M1, which was so large it distorted my leading indices for the US.   I tweaked my "medium" leading index by removing M1 for months when it was distorted, and the spike in the index in 2021 was reduced, though it still rose sharply as fiscal and monetary stimulus during and after the Covid Crash was substantial.

The updated chart with the revised leading index is shown below.  While my "medium" leading index suggests a slowdown, it doesn't yet point to a deep recession, which, nevertheless, I think very likely.

See previous chart/index here.

"Medium" leading index points to 12 months of decline.

 I have three US leading indices.  A "short" one, which leads by 3 months, but sometimes by a bit more.  A "long" one, which leads by 18-24 months, giving you long advance warning of recessions.  And a--for want of a better word--"medium" leading index, which leads the business cycle by ±12 months.  In the chart below, this-"medium" leading index is plotted with a 12-month lag against my US QCI, which is a monthly proxy for quarterly GDP.   (It's not a perfect fit with GDP, but the fit is surprisingly close.    The QCI is an unweighted average of industrial production, retail sales volume, and payrolls employment.)  Relative to trend,  the QCI/GDP has prolly peaked.

The leading index is distorted by the Fed's change in regulations about interest-bearing accounts, which affected the measurement of money supply.  The rise in the leading index in 2021 is therefore prolly overstated.

So far, the "medium"-leading index suggests a downturn, but not a deep one.  But if you reduce the peak in the leading index for the distorted money supply data, the low is the leading index must also be reduced.  I'm doing a bit of research on the money supply data, and also on the leading index if we leave out money supply. [Update: 19/10/22 See updated chart here]

As usual, data have been extreme-adjusted by the Bureau of Census's algorithm (my program)