Friday, December 28, 2018

Marking my 2018 forecast to market

At the beginning of every year, I give a forecast for each of the two halves of the year. And at the end of each year, I look back and mark my forecasts to market. In that way I try to be completely transparent, and to keep myself honest.  Since we're at the end of 2018, let's look back to last January and see how I did.

This year the call is pretty easy.  With the exception of a hurricane-induced whipsaw in September and October, for the last year the L.E.I. [Index of Leading Economic Indicators] has improved by about .4% a month. This strongly suggests clear sailing in the first half of 2018.
....  if you wanted an even easier, quick and dirty approach to a short term forecast, you can simply chart weekly initial jobless claims and the S&P 500 (both of which are components of the LEI).  If both of those are still making new lows/highs respectively (and they have been), then the economy should be in good shape for the next few months. 
....They're good and so is the near term economy.
As I wrote at the time, that was a pretty easy call, because there had been a surge last autumn of almost all of the leading indicators. 

Turning to the second half of 2018, last January after reviewing all of the long leading indicators, I wrote:
To summarize the results: 
  • There is only one outright negative [long leading indicator], and mixed one at that: Corporate bond yields and mortgage interest rates.
  • There are five positives: real M1, the yield curve, credit conditions, corporate profits, and real retail sales per capita.
  • Housing is mixed, neutral to mildly negative for most of the year, but a positive late in the year. Real M2 is also positive for this year, having just turned negative.
My sense is that later 2018 will be weaker than 2017, as the housing market stagnation of Q2 and Q3 2017 feeds through into the coincident indicators. On the other hand, the clear majority of long leading indicators remain positive. Left to its own devices, I do not see any recession for the economy at any point in 2018. 
So now let's take a look at the biggest coincident indicators for the economy.  First, real GDP:

Next, Industrial production:

and finally, employment:

A year ago I thought the forecast for this year was pretty obvious: there would be no recession. And there wasn't.

Turning to the forecast that the first half would be stronger than the second half, Q1 GDP wasn't as strong as 2017 GDP Q2 and Q3 accelerated considerably. So in terms of GDP, the forecast for strength took one quarter longer than expected. Further, both industrial production and employment picked up in the first 8 months of the year (averaging +0.5% for production, and 200,000 jobs) before decelerating beginning in September. So by those two metrics, the forecast was pretty much spot on.

In summary, I think I did a pretty good job forecasting 2018.  I am expecting the 2019 forecast, which I'll post in two parts in January, to be considerably more challenging.