11 forecasters in a Wall Street Journal survey rate the chances of a US recession in the next 12 months at 50% or higher. Since economists almost never predict recessions we can confidently assert that either a contraction has already started or – more likely – will be avoided.

US GDP grew at a 4.0% annualised rate in the second quarter and is likely to register further expansion in the third. Despite housing gloom consumers continue to spend: real outlays rose in July and available evidence suggests a further gain in August. Payrolls growth has slowed sharply in recent months but this partly reflects a surprise cut in government jobs, which is likely to prove temporary. Daily information on withheld tax receipts does not suggest the labour market has fallen off a cliff.

Money market dislocation will hit growth in the fourth quarter but the impact is uncertain and will depend on how long current conditions persist. A useful summary measure of system dysfunction is the spread between the discount rates on non-financial commercial paper and Treasury bills. On three month paper the gap is currently 110 basis points – far above the normal 25-50 b.p. and a level historically consistent with recessions. I will be concerned if this gap fails to narrow significantly by mid October.

The chart shows two versions of our US recession probability indicator, designed to look out six months. The original version suggests recession risk remains low but arguably fails to capture current market dislocation. A new version was therefore estimated including the commercial paper / Treasury bill spread. Assume conservatively that the spread averages 100 b.p. over the remainder of the year. Even on this basis the model suggests the odds are (just) against recession.