In a sense, bond markets came into the week as an underdog.  They'd rallied the entire previous week to the lowest levels in over a month.  And by Friday, that rally momentum was being shed in just such a way that suggested a correction was possible to start the current week. 

While that technically happened (Monday and Tuesday were weaker), it's been a fairly half-hearted effort so far.  In the process, bonds have managed to give off new signals that they might be rethinking the correction that looked like it was underway.

Some of these signals come from various technical studies.  These can be seen in the following chart in the the white-circled areas.  From top to bottom:

1. yields have bounced 2 days in a row on the middle Bollinger Band (21-day moving average)

2. Yesterday's stochastic cross in favor of higher yields will 'uncross' today if rates merely hold flat (more simply: when the purple line is under the green line, it's better for rates)

3. MACD is no longer ascending (more simply: blue bars go up = bad.  Down=good)

4. RSI stalled (More simply: red line in the lower half of the horizontal range = good.  Red line NOT ascending = good).

In short, there was some negative momentum, and now it's showing signs of stalling.  A flat or stronger session would confirm this.

The extent to which bonds respond to the economic data remains to be seen, but if any day this week has a chance to see data move markets, this is it.  Jobless Claims happen every week, but one week each month is more important than the rest because it coincides with the pay period that determines the next NFP number.  Today's is the important one. 

In addition, Housing Starts are out at the same time.  If these two reports happen to be on the same page, it will probably be worth something in terms of trading momentum.  Then at 10am, the Philly Fed Survey is always a solid 2nd tier contributor to the econ data landscape.  In other words, it's always capable of moving markets when it's decidedly better or worse than expected.  Things die down quickly after that, so any momentum that's intact after the data stands a good chance of carrying through the rest of the week.