Soft vs Hard data - Mind the Gap

The gap between survey based indicators (Consumer confidence, small business optimism...) and hard data indicators (Investments, Retail sales, Loans...) is at the widest levels we've seen in a while, indicating some emotional overreaction to future "Trump policies".

Couple of banks: Credit Suisse, Nordea, Morgan Stanley have recently voiced their concern along with Atlanta Fed economists

The Atlanta Fed and the NY Fed's GDP forecasts for Q1 2017 have also diverged significantly (respectively 1% vs 3%).

The Fed of Atlanta focus is on hard data while the Fed of NY model incorporates soft data.

I rebuilt two quick time series as well, for hard data I used the Atlanta Fed GDPnow as a proxy and for the softdata I used an average of the yearly changes for : Consumer confidence, Small businesses optimism and ISM.

The right graph shows a mean reverting time series, testing close to 0 for Hurst exponent* and Dickey-Fuller test *.

As observed, when the spread is above 8 the series tend to mean revert pretty fast, within 2 to 3 months.

So expect surprise indices and soft indicators to print lower in the next quarter.

Finally, for the anecdote, a friend of mine highlighted that most of the enthusiasm is coming from...Republican voters!

* Note that sample is small