US job fears out in the open

In a television interview this morning President Trump has told us there will be “another big job number on Friday”, after the surprisingly strong rebound in employment seen in May and June. However, we are not so sure. Covid-19 containment measures have been re-introduced in numerous states as Governors worry a rising number of cases will overwhelm their healthcare systems. The result is that businesses that had re-opened are being forced to close again, with jobs being lost.

Initial jobless claims have risen for both the third and fourth week of July, while data from Homebase, a payroll scheduling and tracking company, suggests employment in the small business sector has been falling since early July. Worst of all, the Census Bureau has recently been publishing data from its new Household Pulse survey, which reports that after 5.5 million job gains from mid-May to mid-June there were 6.75mn jobs lost between mid-June and mid-July.

Admittedly, the Homebase and Census Bureau’s surveys are brand new and we simply don’t have enough history to be able to pass judgement on their predictive qualities for the “official” jobs number over the longer term. Nonetheless, both were pretty close last month to the gain of 4.8mn published by the Bureau for Labour Statistics. Moreover, even Monday’s ISM manufacturing employment index came in below 50, indicating ongoing job losses in that sector.

Today’s data hasn’t been great either. The ADP private payrolls number for July came in at just 167,000 versus expectations of a 1.2 million gain. However, we would caution that this survey hasn’t had a good track record through this current crisis – look how they revised up their June figure from 2.37 million to 4.31 million to bring it closer in line to the BLS’s 4.8mn figure.

Meanwhile, the non-manufacturing ISM survey’s employment component fell deeper into contraction territory at 42.1 (50 is the break-even level and the further below 50 the greater the contraction). This was despite the activity and new orders components rising very strongly, so it perhaps suggests that firms are not confident this will last.

Consequently, we have very little to back up the President’s assertion that we should be braced for a “big” number. Instead we feel there is significant downside risk to the market’s consensus prediction of a 1.5 million gain. Yes, we have to remember the data is effectively measured mid-month to mid month and there was still some job creation going on in the second half of June, but the evidence from the first half of July is much bleaker. As such we have a forecast of 750,000 gain – but again with a bias clearly to the downside.

We are even more anxious about the prospects for the August jobs number – to be published 4 September – given all of this survey evidence and with the economic effects of activity restrictions increasingly biting. Moreover, high frequency credit card data suggest consumer spending is slowing – see www.tracktherecovery.org. This is likely to get worse given the US$600/week unemployment benefit boost has ended with 30 million people experiencing a massive drop in incomes from this week.

Based on today’s market price action none of this matters – equities are at new highs – and you can understand why they have shrugged it off. In the markets mindset a bad number can generate “good things”. Ahead of an election a poor jobs number would be the most likely catalyst to trigger compromise in Washington that would deliver another meaningful fiscal stimulus next week, while it would also heighten expectations that the Federal Reserve could step in with more stimulus.

However, if we get a fall in August retail sales (published 16 September) as well as a fall in jobs (4 September), which we increasingly expect, this could potentially lead to a more meaningful change in market sentiment on the shape of the economic recovery.