The reliability of initial claims in predicting employment fluctuations depends on the state of the business cycle. While generally very useful in forecasting employment during recessions, very early on in recovery however, they lose their predictive value. This is because initial claims for unemployment are an important measure of layoffs, but changes in overall employment depend on both layoffs and hiring.
Employment can fluctuate for one of three reasons: firms are hiring workers, firms are laying off workers, or workers decide to quit. Claims provide us with a window into the layoff side of the labor market, but in order to paint the whole picture we need to also analyze the other main component- hiring- over the course of the business cycle.
During recessions a strong inverse relationship exists as initial claims rise and hiring receeds. During upturns, however, the systematic relationship between claims and hiring found during recessions virtually disappears, suggesting that layoffs are being driven by factors that differ from those driving hiring decisions. Since hiring overshadows claims and claims move independently of hiring, claims alone cannot tell us much about the overall direction in employment.Thankfully, economic theory does not leave us hanging. Because training new employees can be expensive, firms are often reluctant to fire workers as a way to cut costs and will tend to do so only when no other option is available. During expansions, when hiring rates are high, firms are more likely to adjust to a slowdown in economic activity by hiring fewer workers than by laying off existing workers. In contrast, during recessions, many firms seek to reduce the number of employees on their payrolls. In implementing such cutbacks, these firms will be forced to hire fewer new workers and to lay off part of their existing workforce.
In light of this information we should be looking at hiring to determine where employment will be heading, not just initial claims. So even though initial claims came in under 400,000 all that really means is that firms are laying off less workers versus hiring a ton of new workers. Here is a graph of total private hires (right axis in thousands) vs. a 4-week moving average of initial claims (number on left hand side).
Although hiring has yet to come up significantly it is important to note the data cuts off after october so it maybe misleading. Nonetheless even with 'fresher' data we could still assume that we have a long way to go before our precious unemployment rate is brought down. Another important indicator to look at when deciding whether the employment picture might turn is average weekly hours, but I'm feeling too lazy to put a graph of that up.
All the data for my graph came from FRED (Duh!).
McConnell , Margaret M. “Rethinking the Value of Initial Claims as a Forecasting Tool”. From the Federal Reserve Bank of New York’s Current Issues In Economics and Finance; November 1998, Volume 4/Number 11.