But what I was suggesting is not that short term and long term rates move in lock-step. I know that the yield curve is sometimes negative. But isn't it more likely positive?
Yes, it's positive more than negative, but we're not just looking at a binary situation, we're looking at whether or not history indicates that longer-term rates will reliably increase when the FFR increases.
protagonist wrote:When I see the graph (I think it was your second?) of long term rates minus FFR over time, I don't see randomness. I see a long period of mostly negative points on the Y-axis that pretty much directly corresponded with the long, arguably relatively unusual period of stagflation from the late 60s through early 80s. The rest of the graph is positive on the Y-axis the vast majority of time, and to me, that seems to make common sense.
Again, not binary, so I think you should look at the graph more closely, and think about it some more.
What I see is that it wildly
gyrates from 2.5% or more (steep yield curve) to 0% or less (flat or negative yield curve) throughout the entire period
. Since it doesn't mildly gyrate around any particular value, it tells me that the movement of the 10-year rate is not closely related to movement of the FFR.
I know from having looked before that if we look at short-term and long-term rates from the 10,000 foot view over the entire history available, they do look correlated; i.e., all rates generally moved higher from 1954 to about 1980, and all rates have moved generally lower since. But let's take a closer look at the relationship between the FFR and 10-year CMT. First, instead of looking at the difference, let's just look at both of them on the same chart.
Rather than just looking at the long-term trends, let's look at periods when the FFR moved up or down a lot, and then look at whether or not the 10-year CMT moved up or down with it.
It's easy to see periods during which both rates generally moved in the same direction, such as 1976-1980 (up), and 1989-1992 (down). But it's also not hard to find periods, such as the one I pointed out earlier, during which they moved in opposite directions, or the FFR moved MUCH more than the 10-year Treasury. Here are a few examples:
1973-1974: FFR decreased a lot, 10yr increased
1984-1986: FFR increased a lot, 10yr decreased
1986-1987: FFR decreased a lot, 10yr increased
1992-1996: FFR increased significantly, 10yr decreased
1998-2001: FFR decreased significantly, 10yr increased slightly
2001-2006: FFR increased signficantly, 10yr decreased slightly
protagonist wrote:You don't need lockstep predictability to assess whether a bet is good or bad. You just need greater than 50% predictability to say whether, given enough time, you or the house has the advantage.
We're on the same page in terms of rate increases being more likely than rate decreases if we wait long enough, which is one reason I weight my fixed income more heavily toward direct CDs and less heavily toward bond funds. But because there's a reasonable chance that rates will remain low or go even lower over the next five years, I don't abandon bond funds entirely.
My main point here is that I just don't worry much about the FFR as the cause of increasing rates. The cause of increasing longer term rates will be either a robust economy or inflation. What the Fed does with the FFR will be more of a response to those things than a cause of those things. Read Bernanke's blog post about this if you haven't done so yet.
protagonist wrote:I see, very simply, two points that seem to suggest the future movement of short term rates, and thus also long term rates (if there is any positive correlation). One is the Fed's declaration. The other is the near historic low rates currently (I am not sure I believe in reversion to the mean, or even if we know what the mean is given only a century or so of data, but if it is the natural order of things, that would be a strong signal). I see no strong signs to counterbalance these two arguably strong signals (to a layman like me, anyway). That would indicate more than 50% predictability of the direction of future rates (short, eg. FFR or long term), which would mean if I acted on it, I would have the advantage over the house.
Again, on the same page in terms of historic rates, but not in terms of what the Fed has "declared".
I view the Fed's stated intention to gradually increase the FFR, dependent on the data
, as a statement that they think the economy is in good enough shape that they can start returning to a more normal short-term interest-rate environment. That's great, and I hope the economy continues developing in a way that justifies the rate increases, but I also think that there's much uncertainty about this. And as history shows us, what the Fed does with the FFR does not foreordain what will happen with longer-term rates.
Having said that, I've done some analysis of the correlation between the FFR and the 10-year Treasury, and it is indeed positive on average since 1954, and the rolling average correlation increases as you look at longer rolling time periods. So this supports your position about the odds being in your favor in terms of longer-term rates increasing if the FFR increases. Maybe I'll share some of those results in another post.