An Unexpected Surge

There was an unexpected surge in initial unemployment claims, even as other economic data showed some improvement. Following on the heels of an extremely weak Q1 GDP report (flawed as GDP is as a measure of 'growth'), this continued the recent streak of 'mixed' and seemingly incongruous economic data points. Keep in mind that this Thursday's initial claims report was already outside the reporting period of the payrolls report that will be published on Friday.

As Lee Adler reports here, actual, unadjusted claims confirmed the surge in the seasonally adjusted number this time. Often the two data series fail to confirm each other, and as Adler frequently points out, the statistical 'smoothing' exercises usually only serve to obscure what is really happening. This week's number was quite bad, which is why it is worth mentioning. According to Adler:

“The actual weekly change of an increase of 18,000 compared with the 10 year average for this week of a decline of -12,200.In this week last year the week to week change was a drop of -24,700. In that respect as well, the current reading was bad.”

Calendar quirks (due to Good Friday) may have influenced the number, so one will need to watch for upcoming releases to see if a worsening trend is developing. What is already certain is that the unadjusted claims data have begun to diverge from the stock market.

Initial and continuing unemployment claims. These are the officially reported seasonally adjusted figures – click to enlarge.

It has been observed that the trend in initial unemployment claims and the stock market correlate very closely. Moreover, there is usually a disconnect developing between these trends both at important peaks and important lows. Here is a long term chart showing the inverse of initial claims vs. the SPX. The vertical lines are aligned with major stock market peaks and lows (the last one is obviously only aligned with the 'high so far', since the market may move still higher).

Initial claims (inverted) and the S&P 500 Index -  a close correlation. Divergences (slight leads and lags) can be observed near turning points. Note the subtle deterioration in claims during 2007, while the stock market surged to a new high in late 2007. At the 2009 low, the market bottomed two weeks ahead of initial claims – click to enlarge.

This week's spike in claims may yet turn out to be a one-time outlier – there have been several of those before – but it is a data point that needs to be watched closely, as it usually gives advance warning when the stock market approaches a medium to long term peak (it seems that it lags at lows, so it is apparently not useful for picking bottoms).

The Bond Market's Message

When economic data are mixed, market participants are free to believe whatever they want about the state of the economy. The glass can be seen as either half full or half empty, and at times different markets will disagree on this point.

It is interesting that the 30 year bond yield topped out one precisely one day before the Fed's 'QE tapering' officially started. This actually mirrors bond market behavior that could be observed previously: during 'QE1' and 'QE2' bond yields actually increased while the Fed engaged in debt monetization, as inflation expectations rose. Once the 'QE' exercises concluded, bond yields promptly began to decline, along with inflation expectations. In the course of 'Operation Twist' and 'QE3', this correlation initially broke down for a while, but then reasserted itself in the latter stages of 'QE3'.

The 10 year note yield and the 30 year bond yield. They peaked right at year end 2013, along with the Nikkei and the DJIA (the DJIA has made a marginal new high in the meantime) – click to enlarge.

In all likelihood the bond market is once again telegraphing that the slowdown in monetary pumping will eventually bring about a slowdown in economic activity, similar to what happened after 'QE1' and 'QE2'. Note that the stock market tends to lag the bond market on these occasions.

Conclusion:

One should keep a close eye on both the claims data and the 10 year note. If the 10 year note yield breaks below lateral support and follows 30 year bond yields lower – an outcome that seems highly likely given the frequent testing of said support – it is hard to imagine that the party in the stock market will remain unaffected. 

Charts by: stockcharts