Jobless claims: the Thursday release with a small bark

[Fairman]

Jobless claims: the Thursday release with a small bark

US weekly jobless claims come out every Thursday at 8:30am New York. The headline is a single number, and it's rarely a big surprise.

But it matters in a specific way: it's the most frequent data point on the US labor market, and in periods when the labor market is the main driver of Fed expectations, it moves USD pairs.

The effect is usually small, 10 to 20 pips. The spread widens for a few seconds. The noise level is high, with the first move reversing as often as it continues.

Because it comes every week, traders become complacent. The release that looks routine suddenly produces a 40 pip move when it deviates significantly from the trend, and those who treated it as noise get caught.

The relevant skill is to know when a release matters. During a stable period, it's a non-event. During a shift in the labor market, it's a key indicator.

A rule of thumb: if the last two weeks' claims figures moved the market, the next will probably do so too.

Last month, how much did USD pairs actually react to jobless claims?

[Fairman]

US initial jobless claims are released every Thursday at 8:30am Eastern. They measure the number of people filing for unemployment benefits for the first time, giving a weekly snapshot of the labour market. The release usually has a smaller impact than the monthly payrolls report, but it can move the dollar if the number is far from expectations or shows a trend change. During periods when the market is focused on labour market weakness, claims can matter more. The data is often released alongside other reports on Thursdays, which can make the reaction harder to attribute to claims alone.

[Amanhasnoname]

I looked at this on my own charts over two months of Thursdays and the average move was smaller than most people assume, around 12 pips on EURUSD in the first minute. But it varied wildly. On weeks where the claims figure was within a few thousand of consensus, the move was often under 8 pips and faded fast. On two occasions, the number missed by around 20 thousand and the pair went over 35 pips in a few minutes.

The rule of thumb in the original post matches what I saw. When the previous two weeks had moved the dollar, the next one moved it as well, and when they hadn't, I got a lot of false starts. I now note the reaction size each week in a small table, and if the last two were under ten pips I skip the release entirely.

I also noticed that the spread on my account widened to around 2 pips for several seconds, so waiting about thirty seconds before entering saved me more than any filter did.