WITH Good Friday marking the start of the Easter holiday, investors could be forgiven for expecting the shortened week to lead to traders’ indifference and an “end-of-term” atmosphere. Yet this could be a crucial period for markets. The first quarter comes to an end on Wednesday, and with it the Fed’s $1.25 trillion mortgage-backed securities purchase programme ceases. US Treasury yields have been pushing higher and last week’s auctions didn’t go as smoothly as hoped. To round things off, Friday sees the release of the all-important US non-farm payroll data for March. Unusually, most markets will be closed.
On balance, analysts expect a month-on-month increase in payrolls. If they are right, this would be only the second rise since December 2007. The average call is for an increase of around 200,000 jobs, although estimates range from a small decrease to gains of 400,000. However, the devil will be in the detail. Temporary hires for the US census could add more than 100,000 to payrolls and, as many point out, we just don’t know how many discouraged workers there are: ie, individuals who have given up looking for work and no longer count as unemployed. The Labor Department’s business birth/death model has also added hundreds of thousands of “jobs” by estimating that, despite the recession, there have been more small business start-ups than failures. Whether this is correct or not, it is counterintuitive and there is a danger of future downward revisions.
Despite coming in slightly better than expected last Thursday, the weekly unemployment data continues to disappoint. What looked like a definite change in trend last year now looks ambiguous. Also, while job creation has been evident in the public sector, that too could change. Collapsing tax revenues means that a number of individual states, including California, are unable to balance their budgets. Without Federal help, job losses look inevitable.
The two major concerns this week will be bond yields and payrolls. Expectations are high for a big rebound in the jobs number, but be wary of reading too much into one piece of data.