Earlier this month, I did an interview with the San Francisco Chronicle about repeat no-cost refinancing in a declining rate environment. The piece came out January 4, the day after minutes from the Fed's last 2012 rate meeting were released. Those minutes revealed more bias toward ending Fed support of low mortgage rates than previously thought, and rates rose immediately.

So is the free refi boom over? Not quite, but we're getting close. Below I explain, first by defining "free refi," then by offering some rate market context.

In the Chronicle piece, I explained how a smart repeat refinancing strategy is to do no-cost refinances while rates are dropping, then when rates are at the bottom, doing a normal-cost refinance (which has a rate that's .125% to .25% lower than a no-cost refinance) or even pay points to buy a rate down to capture the true lowest possible low. Here's an excerpt:

Likewise, Julian Hebron, vice president of RPM Mortgage in San Francisco, said, "As rates continue to drop, refinancing repeatedly is quite common in the past 24 months."

Many clients chose no-cost refinances, in which lenders pay all closing costs in exchange for the borrower taking a rate that is one-eighth or one-quarter point higher than the current market rate, he said.

"No-cost refinances are the best play in a declining-rate environment," Hebron said. That's because they allow borrowers to refinance multiple times without digging into their own pockets.

"We've been advising clients not to pay points for most of 2012, on the belief that rates are remaining at existing lows or dropping," he said.

Up until now, this no-cost refinance strategy has worked for borrowers nationwide. But the Fed minutes were the first reminder to markets and consumers that rates won't stay this low forever.

The December 12 Fed meeting statement left markets thinking the Fed would keep overnight rates near zero and continue their agressive mortgage bond (aka MBS) buying to keep mortgage rates ultra low until the unemployment rate drops from current 7.8% to 6.5%. After that, rates were in true record low territory for the rest of December-the perfect environment for the no-cost refi, and most lenders were running full tilt through the holidays to capture and close these loans for borrowers. But when markets got the detailed minutes from that meeting on January 3, it was a different story: "several" members advocated to slow or stop MBS buying (aka QE3) by the end of 2013.

Initially rates rose by .125% on this news. As noted above, a free refi is accomplished by the consumer taking a rate that's .125% to .25% higher in order for the lender to credit the closing costs. So this early-2013 rate spike eliminated that free refi option until this week started. MBS markets rebounded Monday through Wednesday, rates dipped, and were back to no-cost refis at record lows.

But Thursday was a wildly different story as MBS markets sold off sharply (FNMA 30yr 3% coupon down 45 basis points to 103.94), causing rates to rise again. My friend Matt Graham, lead MBS analyst at Mortgage News Daily, explained it well before U.S. data hit Thursday:

On Tuesday Japan's Economy Minister Amari threw a bit of cold water on the perception of a rampantly weakening Yen, but this morning has said that a weaker yen shouldn't be much of an issue until it his 100.0. It was trading as low as 88 yesterday and it now up to 89.44 following those comments.

Amari's comments along only account for a portion of that movement however. They were most damaging as the catalyst for a snowball in European markets leading toward higher rates. That merely started the snowball rolling. Europe gave it a bigger push.

Rumors that the ECB is considering tighter collateral requirements, which sparked concerns about paybacks on the ECB's LTRO's (Long term refinancing operations) at the end of the month. Think of that like a "reverse QE." Just like when the Fed pumps money into the system to lower interest rates and encourage lending, if European banks are pumping money back out of the system (by paying more of it back to the ECB), the implication is for higher rates.

German Bunds moved from 1.47 to 1.55 by 6am New York time. That's a big move! Treasuries were following in relative lock-step, but as is typical during European hours, the movements in Treasuries were an order of magnitude smaller than Bunds, moving up only 5.5 bps vs the 8bp move in Bunds.

That was cause for celebration as it seemed that the massive, unexpected overnight weakness was already in the process of bouncing off a ceiling at 1.86%. Fannie 3.0 MBS had opened at 104-10-a supportive floor earlier in the week-and were hoping to hold there, or close to it. Domestic bond markets were certainly on the back foot, but with some hope of a bounce back with the help of cooperative U.S. data.

But Thursday's U.S. economic data was much stronger-lowest jobless claims in 5yrs, and new home construction at 4yr high-which kept bonds lower and rates higher through the close.

Net result: we end a wild week with rates up .125% again.

Which means rates are still near record lows if you're doing a normal-cost refinance, but a no-cost refi will be higher than it was during late 2012 (and higher than any quote you may have gotten from your lender early this week).

So where do we go from here?

There's a case to be made that rates won't spike sharply because mediocre global economicfundamentals and fiscal paralysis in the U.S. and Europe generally bode ok for bond markets and rates. And if we look back to those Fed minutes indicating Fed MBS buying may end this year, some disagree, including Goldman Sachs:

So what should we make of the FOMC minutes, which suggest that most Fed officials expect to end QE3 by late 2013? Not too much, in our view. For one thing, it is important to remember that the outlook for monetary policy depends on the outlook for the economy. The midpoint of the committee's "central tendency" forecast for real GDP growth in 2013 is 2.65%, which probably implies growth of 3-3½% in H2 given the obvious headwinds in H1. If that turns out to be too optimistic, as we suspect it will, QE3 will probably last longer than Fed officials currently expect. More importantly, the minutes have a tendency to mislead at times when the range of views on the FOMC is large because they paint an overly "democratic" picture of the decisionmaking process. Even under Ben Bernanke-a much less autocratic chairman than many of his predecessors-it is ultimately the Fed leadership that drives the decisions, and it is their views that we need to identify. This is difficult to do with confidence in the minutes, but we suspect that it was mainly the leadership that "...emphasized the need for considerable policy accommodation but did not state a specific time frame or total for purchases." Based on our own expectations for the economy and our understanding of the reaction function, we continue to expect that QE3 will run through 2013 and-at a reduced pace-2014 as well.

That said, it's not just about waiting for the Fed to ease off this multi-year MBS trade that began when QE1 MBS buying was announced in November 24, 2008. The Fed didn't actually buy any MBS until January 2009, but in those weeks from Thanksgiving 2008 to New Years 2009, rates dropped a whopping 1.25% as investors piled into MBS ahead of the Fed. Bond king Bill Gross said at the time:

PIMCO's view is simple: shake hands with the government; make them your partner by acknowledging that their checkbook represents the largest and most potent source of buying power in 2009 and beyond. Anticipate, then buy what they buy, only do it first: agency-backed mortgages, bank preferred stocks, and senior bank debt; Aaa asset-backed securities such as credit card, student loan, and auto receivables. These have been well-advertised PIMCO strategies over the past 6 months but there are others in clear sight.

The operative words here are "do it first." Just like global investors piled into MBS in 2008 and caused rates to drop 1.25% before the Fed spend one dime on rate stimulus in 2009, the exit will work the same: MBS markets will start selling long before the Fed formally indicates their exit, and rates will spike accordingly.

All of this is a long way of offering a wake up call to rate shoppers "holding out for better" because the best levels have been tested many times, and better is unlikely.

It's also a reminder to those who've done a few no-cost refis over the past couple years to do one last check to see if it's mathematically sound to do one final fix of your rate before the market turns.