RateWatch #434 Only Three Things Move Rates

November 20, 2004 by Dick Lepre

Only Three Things Move Rates

Since this past week had some very interesting inflation numbers, it would serve well to discuss
what it is that affects Treasury yields and, consequently, mortgage rates.

There are, I think, only three things that determine Treasury yields and mortgage rates:

1) the perception of inflation
2) the technical factors that we allude to when we refer to StoMaster
3) at those times when rates fall a great deal, convexity hedging is also a factor.

Treasury debt and mortgage-backed securities are purchased by a set of people apart from
us normal working folks. The average person is more likely to invest his savings in equities
because, while riskier, they offer a greater return on investments.

Holders of fixed-income securities (Treasury debt, corporate bonds and mortgage-backed securities) have a different purpose. They are already wealthy. They do not necessarily want to accrue more wealth, their investment strategy is to preserve the purchasing value of their present wealth. For them, money is not a medium of exchange but a manner of storing present wealth for future spending. For them, there is only one enemy - inflation. Inflation means that their money will not buy as much later as it will now. As they park their wealth in fixed-income securities they demand a return that will offset the effect of inflation. If inflation moves up they will not purchase fixed income securities unless the yield on those securities rises to offset the effect of inflation.

Therefore the first rule that sets mortgage rates is the perception of inflation.  Note that I keep saying "perception".  If inflation is happening then, most assuredly, rates will rise.  It inflation is perceived to be "about to happen" that is good enough to move yields up.

The things that move yields up are the economic releases that come out weekly, monthly and quarterly.

The quarterly report is GDP. There is a preliminary release of GDP and two revisions. The variance in the revisions is largely the effect of imports and exports.  The preliminary number is based largely on Retail Sales.  If more of those retail sales than expected were imported items then GDP (Gross Domestic Product) will be smaller.

The general belief is that high GDP growth sets the stage for inflation.  This is, to some extent, a relic of classical Keynesian economics - GDP means increased demand which is per se inflationary.  The experience of the 1990s served to inform us that expansion can take place without anything but modest inflation.  It was probably the case that a widening of the global economy which meant more supply is what kept inflation low and continues to do so.

Actual inflation numbers: CPI and PPI report real inflation but are often the source of
confusion.  PPI is inflation in wholesale prices.  If those increased costs can be absorbed because the market is competitive and profit is sacrificed for market share then those increased wholesale costs may not be passed along to the consumer.

Further confusion results from the "overall" vs. "core" numbers that come out each month.  "Overall" means just what it says - the price of a predefined large "basket" of items.  "Core" recognizes that two things: energy and food have large month-to-month fluctuations that may be
a function of things such as weather and other supply-limiting effects such as strikes and that the spikes in these two areas should be filtered out.  Energy prices are particularly annoying because they affect the cost of many other things.  Eliminating the effect of high energy prices and looking at the "core" number amounts to saying, "We are only going to concern ourselves when the effect of increased energy costs is seen elsewhere."

From my perspective, the most important inflation data each month is core-CPI.

Strangely enough, the surrogate for CPI has become the monthly BLS Employment Situation Report.  People still act as if this is an accurate forecast for inflation which it is not.  It may be the case that since the headline jobs number is extremely difficult to predict the large extent to which it came come is different from expectation creates a "surprise effect" which moves the market.

The second factor that affects rates is the technicals.  We have discussed this extensively in the past.  There are cycles in bond prices and, consequently, mortgage rates which are caused by nothing more than the patterns caused by buyers and sellers.  People get into and out of positions for speculative rather than long-term reasons and there is a pattern to this behavior that we discuss often as we analyze the stochastic technical.  The technicals say more about the people doing the trading than they do about the commodity itself.

For example, at present, we are in a downcycle (lower prices, higher yields) for the week-to-week tech which should last until mid-January.  The longer term tech is bullish. Think of the two periods of interest the way you would think of the variation in temperature. The longer term tech is like the differencein temperature from winter to summer.  The shorter term tech is like the difference in temperature from night to day.

The third effect is an occasional one. Mortgage refinancing can itself cause rates to fall. This effect is called "convexity hedging". This occurs when large numbers of mortgages are paid off early and investors find that duration of their assets does not match the duration of their liabilities. At such a time, they may buy 10-year Treasury notes or 30-year bonds merely to increase the average duration of their assets.  This happened during the last bull cycle causing interest
rates to fall to record lows. 

To receive this free RateWatch newsletter each week, sign up here.
RateWatch Archive