So how good is APD at predicting NFP, anyway?

183k jobs were added to the US economy in January, according to the ADP employment report. And that is a solid 25k above the 148 forecast, or 61k above the initial estimate of 122k. December’s figure was also upwardly revised by 64k to 176k, making it a trifecta of positive metrics for the headline figure alone.

 

With such a solid set of numbers, one cannot help but wonder if this will have a positive knock-on effect for January’s NFP numbers released on Friday. They tend to trend in the same direction overall, but I wanted to see how they compared on a month-by-month basis. So I decided to take a look at the predictive quality that ADP may have on the headline Nonfarm payrolls figure. I wish I hadn’t.

 

 

Can ADP’s direction (current v prior) indicate the direction of NFP?

I’m not looking for ADP to nail the NFP number, as I believe that to be an unrealistic task given the different methods used to collect data for each report. But perhaps a month-over-month rise for ADP can indicate a hotter NFP (or weaker ADP hinting at a softer Nonfarm figure).

 

• Data is 2007 to 2024
• Initial (unrevised) figures are used, as this is what markets react to
• A ‘rolling success rate’ is show to display the trend of its predictability over time

 

 

There is no escaping that fact that flipping a coin would do a better job than ADP predicting NFP’s monthly direction, with a success rate of just 47.7%. That means it was incorrect 52.3% of the time. If one was to use this for anything at all, it would be to fade the move. But even then, it is too close to 50% for most to successfully use as part of an edge, especially when we also need to factor trade entry, trade management and exit.

 

The low success rate also explains why were have a clear downtrend over the sample period, and why we’re likely better assessing the overall trend of ADP alongside NFP, and not using it as a predictive tool on a month-by-month basis.

 

 

Now onto Nonfarm payrolls

I may not have found what I wanted to see, but we can still leave this on a positive. The consensus forecast for NFP this Friday is for employment growth to dip to 154k from 254k. And with no help from ADP, I find this hard to believe, given NFP tendency to beat forecasts roughly 67% of the time over the last 12 months, and 5-year windows. Alongside the strong economic data coming out of the US in general.

 

 

 

-- Written by Matt Simpson

Follow Matt on Twitter @cLeverEdge