Deprecated: Assigning the return value of new by reference is deprecated in /f2/blendedtechnologies/public/wp-content/plugins/pmetrics.php on line 1192
Blended Technologies » Blog Archive » There’s a 60.9% Chance Interest Rates Won’t Change Next Month - Bayesian Classification of Fed Statements

There’s a 60.9% Chance Interest Rates Won’t Change Next Month - Bayesian Classification of Fed Statements

I remember hearing in an Economics class long ago that the Federal Reserve has to give hints to the public about future interest rate changes since surprising financial markets is usually a bad thing.

I got to wondering recently if there were certain key words the Fed might use in its statements to warn the markets about upcoming changes in rates.

Coincidentally I also recently came across a very easy to use Python based Bayesian classifier called Reverend. (Here and here are some good introductions to the concept of Bayesian classification.)

So I came up with a plan put all of the FOMC statements I could locate into a Bayesian classifier and see if I could use words in past statements to predict future rate changes.

Latest (and better) code with more recent FOMC statements here.

Here is the data I compiled:
FOMC Statements with Rates (Includes two files, a Gnumeric format and a wierd CSV format)

Next, I put together a little Python script to read in the data I compiled, and train two Bayesian models.
Here’s everything you need to run the program for yourself.

One model will try to predict the exact rate change e.g., -.5, -.25, 0, .25, or .5 percent.
The other model tries to predict only the direction (or no change) of an interest rate change.

To test these models, I used 10 fold cross validation on a randomly shuffled data set and then I did 10 runs of that.

From these runs:
The direction-only model succesfully predicted the next rate change direction an average of 72.17% of the time with a standard deviation of 18.8 across all of the folds for all of the runs.
The model to predict the exact rate change was correct an average of 69.5 % of the time with a standard deviation of 17.28.
I don’t have any other methods to compare these results to so I’m not sure if they are good or not. But it looks like the predictions are correct a fair amount of the time, so perhaps this method is at least plausible.

What words in FOMC statements signal rate changes?
I’ll pull out a few of the high probability words for somes of the classes (rate change amounts) so you can have a look.

Lowered rate by .5% at next meeting:
(mostly words associated with a poor economy, pretty neat!)

Raised rate by .5% at next meeting:

(Maybe words associated with an overheating economy?)

Left rate unchanged:

And finally, my predictions for interest rates for July, 2007:

Both models predict a 60.99% chance of no change to the Fed funds rate.

Here are all of the two models’ predictions:

Exact Rate Change model predicts:
‘0′, 0.60992435592948235
‘0.25′, 0.2681162694505409
‘-0.5′, 0.022171742716859288
‘-0.25′, 0.004682228470558325
‘0.5′, 0.00040881787835661543

Direction only model predicts:
‘No Change’, 0.60992435592948235
‘Increase’, 0.28084524515677251
‘Decrease’, 0.13120882568327658 |  Digg |  FURL |  Yahoo! My Web 2.0 |  Reddit

12 Responses to “There’s a 60.9% Chance Interest Rates Won’t Change Next Month - Bayesian Classification of Fed Statements”

  1. David Avraamides Says:

    This is quite interesting. It would also be interesting to look at the timing of the statements compared with the actual market prices for interest rates and related derivatives to see if traders following these FOMC statements interpret them similarly.

    In other words, does this present a trading opportunity or is the information priced into the market fairly efficiently as soon as the statements are relased?

  2. Greg - CEO/Founder Says:

    David, that’s an excellent idea. The markets really are the best predictors for interest rate changes (AFAIK though I’d be curious to see how accurate they really are.) So, in theory, if I look at how the market changes right after the release of a statement, perhaps I could compare my estimate from the latest statement to the market’s interpretation of the statement.

    Nice website BTW. How did you make the styling so nice? You should probably have a “more” link on each blog post on the main page, it’s hard to tell that there’s more to the post than what’s shown there.

  3. efm Says:

    I added the text of the statement today and got this prediction:

    Prediction for rate change:
    [(’0′, 0.60716369942518311), (’0.25′, 0.27171641561489529), (’-0.5′, 0.11
    301353351197091), (’-0.25′, 0.0043705265077812472), (’0.5′, 0.00041987468
    Prediction for if rate will change:
    [(’No Change’, 0.60716369942518311), (’Increase’, 0.28468974884903231), (
    ‘Decrease’, 0.18648144709131648)]

    So, the August meeting is likely to not change the rates.


  4. Greg - CEO/Founder Says:

    Excellent work, Efm! I’ll have to try it myself. This seems like a plausible prediction to me.

  5. efm Says:

    With the August meeting statment I get this prediction:
    Prediction for rate change:
    [(’0′, 0.6021048139175551), (’0.25′, 0.23475207415927402), (’-0.5′, 0.11085195796016573), (’-0.25′, 0.0047277966182899811), (’0.5′, 0.00048592235452094013)]
    Prediction for if rate will change:
    [(’No Change’, 0.6021048139175551), (’Increase’, 0.2487824837342843), (’Decrease’, 0.18332179934596032)]
    So the chance of no change has very slightly decreased, indicating they are not likely to change the rates in the September meeting either.

  6. Greg - CEO/Founder Says:

    Thanks for the quick work on that, efm. I always forget when the meetings are coming. From reading this statement, the predictions here do look plausible. And I the last two predictions have been correct. That’s encouraging.

  7. efm Says:

    I apologize for not getting to this earlier. With the unexpected decrease in rate at the September meeting, the likelihood of no change at the October meeting has dropped by 15% and the likelihood of a .25 increase has increased slightly, and the likelihood of a .5 decrease has also increased.

    Prediction for rate change:
    [(’0′, 0.44746044289510639), (’0.25′, 0.34803196987274321), (’-0.5′, 0.18882184654358186), (’-0.25′, 0.0058650337700004163), (’0.5′, 0.00049887591533603892)]
    Prediction for if rate will change:
    [(’No Change’, 0.44746044289510639), (’Increase’, 0.37159814929233181), (’Decrease’, 0.24615516857981962)]

    I’d have to say anything could happen at the October meeting.

  8. Greg - CEO/Founder Says:

    Thanks, Efm. I agree, it sounds like anything could happen. Keep in mind the model should be more accurate if we update it with all of the Fed statements since I originally wrote the script. I’ll try to do that this weekend.

  9. Ames Tiedeman Says:

    In the U.S. interest rate are going lower, Gold is going higher, Oil is going higher, inflation is going higher, the dollar is going lower. What is wrong with this? Everything! At some point the FED is going to have to raise rates bigtime. We are in a very, very, precarious situation at the moment. I think Gold will tripple to over $2,000 an ounce when the market finally wakes up and sees the real inflation. Last I checked a lower dollar = higher import prices. There is no inlfation deflator here. With commoditioes on fire you can forget about that. Bernanke should have never lowered rates last week. However, the Fed might be doing something that few have talked about. Maybe the Fed has abandoned the dollar the crush teh trade deficit. Good luck, it will take 20 years to correct our 6% of GDP trade deficit and move it back to under 1% of GDP, unless you want to seriously disrupt the global economy. We are in for tough times people. Very tough!

  10. Evelyn Says:

    The forecast, after adding the October Statement (which lowered rates an additional .25% to 4.5% is:

    Prediction for rate change:
    [(’0′, 0.45220429695537889), (’0.25′, 0.39520799597082201), (’-0.5′, 0.1873814776692021), (’-0.25′, 0.0048850804754397292), (’0.5′, 0.00044497458812392532)]
    Prediction for if rate will change:
    [(’No Change’, 0.45220429695537889), (’Increase’, 0.41304520169354481), (’Decrease’, 0.20332618490924276)]

    The decrease this month, although widely expected according to the press, was not predicted by the model last month. Though, I did say anything could happen.

    The likelihood of no change is about the same as last month, that of an either an increase or a decrease have both gone up. After two decreases in a row, I’d bet on no change. The Fed doesn’t like to give equivocal signals.

  11. Greg - CEO/Founder Says:

    Yeah, the model was pretty off for this month. Did you use the latest model for these predictions?

  12. evelyn Says:

    No, I hadn’t updated. I’ll give that one a try. Thanks!