Written by Steven Hansen
The Empire State Manufacturing Survey moved barely into expansion territory in December 2013.
- Expectation was for a reading of 5.0 versus the 1.0 reported
- New orders sub-index of the Empire State Manufacturing Survey improved but remains in contraction, whilst unfilled orders contracted even more.
- This noisy index has moved from -8.1 (December 2012), -7.8 (January 2013). 10.0 (February) , 9.2 (March), 3.1 (April), -1.4 (May), 7.8 (June), 9.5 (July), 8.2 (August), 6.3 (September), 1.5 (October), -2.2 (November), – and now +1.0.
As this index is very noisy, it is hard to understand what these massive moves up or down mean – however this regional manufacturing survey is normally one of the most pessimistic.
Econintersect reminds you that this is a survey (a quantification of opinion). Please see caveats at the end of this post. However, sometimes it is better not to look to deeply into the details of a noisy survey as just the overview is all you need to know.
From the report:
The December 2013 Empire State Manufacturing Survey indicates that manufacturing conditions were flat for New York manufacturers. The general business conditions index rose three points but, at 1.0, indicated that activity changed little over the month. The new orders index inched up, but remained negative at -3.5, while the shipments index rose to 7.7. The unfilled orders index fell to -24.1, and the inventories index declined twenty points to -21.7; both indexes reached their lowest levels since 2009. The prices paid index was little changed at 15.7, and the prices received index climbed to 3.6. Labor market conditions remained weak, with the index for number of employees holding at 0.0 for a second month in a row and the average workweek index dropping six points to -10.8. Indexes for the six-month outlook generally conveyed a fair degree of optimism about future conditions, though to a lesser extent than in the November survey.
This month’s supplementary questions asked manufacturers to assess how much of a problem certain business issues were for their firms and whether the issues were expected to become more or less of a problem in the year ahead. As in earlier surveys, the issue cited most frequently, by far, as a major problem was the cost of employee benefits. Moreover, fully 80 percent of respondents expected that this would become even more of a problem a year from now. Finding qualified workers emerged as the second most widespread problem, eliciting a considerably larger degree of concern than in earlier surveys. This, too, was expected to become more of a problem in the year ahead by a wide margin. In contrast, the availability, cost, and terms of credit were seen as relatively minor problems that would become even less consequential over the next year. For more details, see the full supplemental report.
The above graphic shows that when the index is in negative territory that is not a signal of a recession: of 5 times in negative territory only one occurred with a recession. Conversely, a positive number is likely to be indicating economic expansion. However, when it does make a correct negative prediction it can be timely. This index was only two months late in going negative after what was eventually determined to be the start of the 2007 recession.
This survey has a lot extra bells and whistles which take attention away from the core questions: (1) are orders and (2) are unfilled orders (backlog) improving? Econintersect emphasizes these two survey points.
Respondents believe the level of unfilled orders (backlog) is trending down (looking at a 3 month rolling average; it has been negative since 2011. Unfilled order contraction can be a signal for a recession.
Holding this and other surveys Econintersect follows accountable for their predictions, the following graph compares the hard data from Industrial Products manufacturing subindex (dark blue bar) and US Census manufacturing shipments (lighter blue bar) to the Empire State Survey (darker green bar).
Comparing Surveys to Hard Data
In the above graphic, hard data is the long bars, and surveys are the short bars. The arrows on the left side are the key to growth or contraction.
Summary of all Federal Reserve Districts Manufacturing:
Richmond Fed (hyperlink to reports):
Kansas Fed (hyperlink to reports):
Dallas Fed (hyperlink to reports):
Philly Fed (hyperlink to reports):
/images/z philly fed1.PNG
New York Fed (hyperlink to reports):
Federal Reserve Industrial Production – Actual Data (hyperlink to report)
Caveats on the use of Empire State Manufacturing Survey:
This is a survey, a quantification of opinion – not facts and data. Surveys lead hard data by weeks to months, and can provide early insight into changing conditions. Econintersect finds they do not necessarily end up being consistent compared to hard economic data that comes later, and can miss economic turning points.
According to Bloomberg:
The Empire State Manufacturing Survey is a monthly survey of manufacturers in New York State conducted by the Federal Reserve Bank of New York. Participants from across the state in a variety of industries respond to a questionnaire and report the change in a variety of indicators from the previous month. Respondents also state the likely direction of these same indicators six months ahead. April 2002 is the first report, although survey data date back to July 2001. Each month, new data will be released and the previous month’s data will be revised slightly. Once per year, all data will undergo a benchmark revision.
This Empire State Survey is very noisy – and has shown recessionary conditions throughout the second half of 2011 – and no recession resulted. Overall, since the end of the 2007 recession – this index has indicated two false recession warnings.
No survey is accurate in projecting employment – and the Empire State Manufacturing Survey is no exception. Although there are some general correlation in trends, month-to-month movements have not correlated with the BLS Service Sector Employment data.
Over time, there is a general correlation with real manufacturing data – but month-to-month conflicts are frequent.