Page 9 - ISQ UK_October 2017

APRIL 2019

Making Sense of Noisy Economic Data

Scott J. Brown, Ph.D., Chief Economist

Economic data is critical to the financial markets. It helps the particular activity it is interested in, so it measures a sample.
to drive earnings expectations and is a key factor in Federal Choosing that sample is a science and the various agencies
Reserve (Fed) policy decisions. However, economic figures generally do an excellent job, but that still means there will be
are noisy and reports often conflict with one another. How some uncertainty in the data.
do we make sense of it all? For example, the monthly change in nonfarm payrolls is reported
accurate to ±115,000. That means if the monthly change is
MANY SOURCES OF U.S. DATA reported as +160,000, there is a 90% chance the true monthly
change is between +45,000 and +275,000.
Unlike in other countries, the responsibility for collecting and
publishing U.S. economic data is spread across several agencies. The other major source of uncertainty is due to seasonal
The Bureau of Economic Analysis reports on Gross Domestic adjustment. There is a significant seasonal pattern in most
Product (GDP), household income, and consumer spending, while unadjusted data. For example, we normally lose about three
the Bureau of Census covers things like retail sales, residential million jobs each January following the end of the holiday
construction, new home sales, and durable goods orders. The shopping season. The government does a good job with seasonal
Bureau of Labor Statistics reports on the job market, but it also adjustment, but it’s difficult to get it exactly right.
publishes the Consumer Price Index (CPI) and other inflation
gauges. The Fed produces the index of industrial production. In Economic data are subject to two kinds of revisions. Figures are
addition to the government figures, there are a variety of private- often revised in the following month, reflecting more complete
sector data sources, including the Institute for Supply Management information. Annual benchmark revisions seek to tie the data
(monthly purchasing managers’ surveys) and the Conference back to more comprehensive sources, such as nonfarm payrolls
Board (consumer confidence and the index of Leading Economic to actual payroll tax receipts.
Indicators). For those using the economic data, uncertainty means one
should take any reported number with a grain of salt. It’s best to
SOURCES OF UNCERTAINTY: STATISTIC look at a three-month average, which reduces much of the noise
SAMPLING AND SEASONAL ADJUSTMENT (but does not eliminate it) and is a better gauge of the underlying
There are two major sources of uncertainty in the economic data. trend.
The first is statistical error. The government can’t observe all of

8