Any change in interest rates will be sprung on the markets if John Williams, Federal Reserve Bank of San Francisco president, gets his way.  While we suspect that any shock value surrounding this particular event would be moot, Williams told CNBC this week that "My personal preference is that we don't have the most telegraphed policy decisions in history, as we did in 2004."  He added "You don't want to make a decision two months or three months in advance."

While it appeared unrelated to William's remark, Wells Fargo Bank just released the first in a series of articles which seek to quantify the expected reaction of a fed funds rate shock on several different markets. The bank cites the example of then Fed chairman Ben Bernanke's comments in the summer of 2013 which caused investors to revise their expectations about the Feds monetary policy and rates and led to the so-called "taper tantrum".

The first article looks at the concept of a surprise and develops a theory as to why surprises are employed rather than simply making changes, outlines different measures for measuring surprises, and looks at the sensitivity of the first of several markets, the dollar exchange rate

A surprise component can offset the effect of market expectations.  In theory, as all publicly available information should be discounted into prices, if macroeconomic news perfectly matched advanced expectations asset prices would not move.  In practice this may not always be the case as uncertainty plays a role and because individual market participants may have acted counter to expectations.  As a control for its study Wells uses the surprise component (the difference of the actual release from expectations) rather than actual release value.

One way of determining the surprise component of a fed funds rate announcement is to determine expectations and compare the actual release to it.  The second method is to look at financial assets themselves to see how their prices, which should already incorporate expectations, change following the announcement then calculate the surprise component from the movement. A third method, a statistical proxy for expectations was discounted completely.

Wells' economists decided that the most appropriate instrument for determining expectations regarding a Federal Open Market Committee (FOMC) policy announcement was the fed funds future contracts which they said were superior at predicting the funds rate for short time horizons.  They plotted the surprise component extracted from those contracts for the time period 1994 to 2014, eventually truncating the data set at the end of 2008.  

They partitioned their dataset to study how fed funds rate surprises affect different markets before and during the crisis, defining the beginning of the crisis as the first time the FOMC reduced interest rates for the cycle, at its September 2007 meeting.  They then studied the sensitivity of the trade weighted dollar to surprises in the rate. 

The interest rate differential between two countries is a large driver of the exchange rate and they anticipated that unexpected moves in the federal funds rate would also drive moves in the dollar around these announcements.  These changes would, on average be associated with a strengthening of the dollar.

Using an event-study approach they regressed the percent change in the dollar index around surprises on the unexpected component of announcements in the fed funds target rate.  An earlier Chicago Fed study had found that the dollar/mark and dollar/yen exchange rates were affected by the unexpected component of a fed funds announcement but the effect was not evident until a large amount of time had passed, something Wells Fargo's analysis could not capture from the daily returns it was using.  "Clearly the relationship," they say, "is more complicated than we initially thought."

Wells Fargo's findings support the Chicago Fed's immediate results, that is that fed funds surprises have an insignificant effect on the dollar initially.  This was at odds with its starting hypothesis that the dollar would strengthen immediately following a surprise.  An alternative explanation could be the signaling effect that is contained in the unexpected component cancels out, on average, the effect caused by interest rate differentials.

It is plausible that news contained in a surprise could cause investors to revise their expectations regarding the domestic economy and, given the size of the U.S. economy, could have large implications on the global economy, which could change investors' risk preferences. The economists use the example of an unexpected rate cut sending a signal the economy is weaker than many thought, causing a flight to the safety of the dollar.  Investigating the subsamples before and during the crisis and checking for asymmetric responses to see if the dollar responded differently to positive versus negative surprises all resulted in finding that the sensitivity of the dollar was insignificant.

In subsequent articles Wells Fargo will look at the effect of surprises on other financial assets including Treasury securities and broad equity indices in order to gain insight on the behavior of financial markets surrounding the impending rate announcement and will attempt to quantify the reaction to surprises, both positive and negative.

Incidentally, Williams acknowledged that keeping fed moves close to the chest could lead to some market swings.  "In a normal economy there is some volatility in markets, that is just a healthy functioning of markets trying to understand and filter what the data means for policy," he said. "It's healthy for the future actions to be uncertain because economic conditions can change."