Download FedViews

Twelfth Federal Reserve District
FedViews
March 10, 2016
Economic Research Department
Federal Reserve Bank of San Francisco
101 Market Street
San Francisco, CA 94105
Also available upon release at
http://www.frbsf.org/economic-research/publications/fedviews/
Daniel Wilson, research adviser at the Federal Reserve Bank of San Francisco, stated his views on the
current economy and the outlook as of March 10, 2016.

Despite volatile financial conditions since the start of the year, incoming economic data in recent
weeks have been fairly positive and point to healthy economic growth. While real GDP growth in
2015 was around 2%, we expect real GDP to grow slightly faster this year and slow slightly to just
below 2% next year.

Similarly, in the labor market, we expect the unemployment rate to fall a bit further this year before
gradually rising back toward our estimate of the natural rate of unemployment, around 5%.

Our expectation that the unemployment rate will continue its decline a bit longer also is bolstered
by the solid gains we have been seeing in nonfarm payroll employment. February saw net gains of
242,000 jobs, which is just slightly above the six-month moving average.

On the inflation front, recent data point to some firming, consistent with our outlook that inflation
will gradually rise toward the Federal Open Market Committee’s stated objective of 2% over the
next two years. This outlook assumes that the negative pressures stemming from the earlier declines
in energy prices and the appreciation of the dollar will gradually dissipate.

Despite the generally positive recent data on economic conditions and inflation, market interest
rates such as yields on 2-year and 10-year Treasuries and 30-year mortgages are down from the
start of the year. These declines are likely to reflect a couple of factors. First, demand for U.S.
assets has increased due to a weakening of economic conditions abroad and lower yields on foreign
sovereign bonds. Second, financial market participants in the United States appear to have become
more pessimistic about economic conditions and inflation or more concerned about downside risks.

Many economic variables, such as monthly employment growth, exhibit large fluctuations due to
seasonal factors. A comparison of seasonally adjusted (SA) and not seasonally adjusted (NSA) data
on the monthly change in nonfarm employment shows that the magnitude of seasonal adjustment
can be enormous. For instance, a typical January sees a net decline of over 2 million jobs. By
comparison, the worst month of job losses in the Great Recession, March 2009, saw a net decline of
just a little over 800,000 jobs, according to the seasonally adjusted data. The large magnitude and
The views expressed are those of the author, with input from the forecasting staff of the Federal Reserve Bank of San Francisco.
They are not intended to represent the views of others within the Bank or within the Federal Reserve System. FedViews generally
appears around the middle of the month. The next FedViews is scheduled to be released on or before April 18, 2016.
variability of these seasonal fluctuations poses a daunting challenge for government statistical
agencies seeking to adjust such data for seasonality.

While factors like holidays and school-year schedules explain much of the seasonal fluctuations in
employment growth, typical weather patterns also play an important role. Similarly, atypical
weather fluctuations from month to month can have a large impact on employment growth, even
after adjusting for seasonality.

By combining comprehensive local weather data from the National Oceanic and Atmospheric
Administration with local employment data from the Bureau of Labor Statistics, one can estimate
exactly how different types of weather affect employment growth. Specifically, one can measure
how much monthly employment growth at the county level varies with weather in that county in the
current and prior months. The analysis reveals that, in general, weather tends to have large but
transitory effects. For example, winter rainfall, snow, and reported storm damages depress
employment growth on impact, but the effects reverse in the following month. A higher frequency
of especially cold days, on the other hand, depresses employment growth in both the current and
subsequent month.

This estimated county-level model can be used to predict how much employment growth in each
county in any given month was boosted or damped by atypical weather, such as colder temperatures
or more snow than normal. Aggregating these weather adjustments across counties gives an
estimate of how much national employment growth was boosted or damped by atypical weather
across the country.

Applying this methodology to the current winter pattern indicates that the strong employment
growth in December 2015 was boosted by weather, while the relatively weaker employment growth
in January 2016 was damped by weather.

This empirical model of local employment growth as a function of weather can be refined in several
ways. For instance, one way is to allow the effects of certain weather variables to be different for
counties in different regions of the country. The key point, however, is that by exploiting the rich
granular data available on local weather, such models hold much promise for helping policymakers
better discern how much the movements in incoming economic data are due to changes in the true
underlying strength of the economy rather than to fluctuations in the weather.
Economy to grow slightly above trend
Real GDP
Percent change from 4 quarters earlier
FRBSF
Forecast
Labor market close to normal
%
Unemployment rate and forecast
4
Sesaonally adjusted monthly observations, forecast is quarterly average
%
12
3
10
2
Q4
1
8
0
Feb.
4.9
-1
6
4
FRBSF
forecast
-2
-3
2
-4
-5
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2008
2009
2010
2011
2012
2013
2014
2015
2016
0
2017
Source: Bureau of Labor Statistics and FRBSF staff
Source: Bureau of Economic Analysis and FRBSF staff
Job growth remains healthy
Inflation to rise gradually to target
Nonfarm payroll employment
Thousands
Seasonally adjusted, monthly change
Personal consumption expenditures (PCE) price inflation
%
Percent change from 4 quarters earlier
5
Monthly
change
300
Feb.
242,000
6-month
moving
average
4
FRBSF
forecasts
Core PCE
price index
200
100
2
1
Q4
Overall PCE
price index
3
0
-1
0
2013
2014
Source: Bureau of Labor Statistics
2015
2008
2016
2009
2010
2011
2012
2013
2014
2015
2016
2017
-2
Source: Bureau of Economic Analysis and FRBSF staff
Since Fed hike, market rates have fallen
Interest rates
Weekly average
Seasonal adjustments are enormous
%
Nonfarm payroll employment, monthly change
8
seasonally adjusted (SA) vs. not seasonally adjusted (NSA)
Millions
1.5
7
0.5
6
5
30-year mortgage
-0.5
03/04
4
-1.5
3
Ten-year Treasury
2
-2.5
SA
Two-year Treasury
Federal
funds
rate
2008
2009
Source: FAME
2010
2011
2012
2013
2014
2015
2016
1
NSA
0 1990
1995
-3.5
2000
Source: Bureau of Labor Statistics
2005
2010
2015
"Weak" January job growth due to weather
Weather has large but transitory effects
Effects of weather on employment growth
%
Winter Months Only, Jan. 1990 - Dec. 2014
Current Month
Prior Month
0.06
Weather-adjusted employment change
Monthly change; seasonally adjusted
0.04
Thousands
Dec. 2015
Jan. 2016
Feb. 2016
300
250
0.02
200
0
-0.02
150
-0.04
100
-0.06
-0.08
Precipitation
Snowfall
Number of days
with minimum
temperature <30
50
Total damages
Notes: Coefficient shows pctg. pt. change in employment growth due to one std. dev. change in weather variable.
Source: Author's calculation, based on data from BLS (QCEW) and NOAA.
0
Official BLS (not weather-adjusted)
Source: BLS and FRBSF
FRBSF (county model)