Alternatives to market moving data during US government shutdown
The shutdown of non-essential federal government functions in the United States has halted data collection, analysis and the publishing of several key economic indicators. Fortunately, the Federal Reserve is classified as an essential institution and remains open, so in addition to the data collected by the private sector, investors are not left completely in the dark.
Even if the shutdown ended right now, the data would not be instantly released. The September employment report data was already collected before the shutdown, so after crunching the numbers for a couple of days, it would probably be the first report to be released. For many other series though, data collection has been completely stopped because of the shutdown, and the delays could be much longer. Such data collection delays increase the odds that the data will be less reliable and more prone to later revisions.
The lack of key economic data during the US government shutdown has prompted economists, investors and traders alike to look at alternative sources to gauge how the economy is doing. Photo source: shutterstock.com
Estimating missing data
It is relatively easy to find close substitutes for some of the delayed data. For other reports though, an estimate has to be developed from closely correlated data. Unfortunately, estimation can become messy and introduces an error risk to the equation. Thus, it makes sense to first decide what is required, and only then think about what is possible.
What most investors want from data is the ability to predict central bank policy and general economic growth. As the Fed is currently data-driven and policy is tied to inflation and unemployment thresholds, at least these two numbers should be estimated. For economists, the shutdown’s negative growth effect and the shock’s half-life are of interest. Unfortunately, this introduces new variables to the estimation. It is hard enough to estimate the missing data, but add to this the probable negative effect of the shutdown and related drop in confidence, and you are really trying to hit a moving target.
Price indices can be estimated from several sources that collect price data. PriceStats collects data from hundreds of online retailers on a daily basis in order to estimate fluctuations in consumer prices at a daily level. Using online data allows reporting with high frequency and it is plausible that online prices are adjusted quicker than other data. PriceStat does not try to forecast the consumer price index, but the recent data shows that prices remained weak until mid-September, and have risen slightly since then.
The increase in payrolls and the unemployment rate are very important numbers, but they are unfortunately volatile and prone to later revisions. One possibility is to look at the changes in weekly jobless claims and use them to forecast the payroll number. The trouble is that while there is a clear negative correlation between weekly claims and the change in payrolls, the explanatory power of claims is quite weak. For a brief discussion and charts, see Financial Sense’s article.
A more refined way is to look at the Institute for Supply Management’s monthly employment indices. Note that in chart I have averaged the manufacturing and non-manufacturing employment indices.
Source: Saxo Bank
Industrial production and inventories can be estimated from the ISM monthly report on manufacturing and its several sub-indices, together with Markit’s PMI-indices. Retail sales volume can be extrapolated with the help of weekly retail sales tracking by the International Council of Shopping Centers, published every Tuesday. The previous report showed that after the shutdown commenced, weekly sales fell 0.1 percent, which decreased the year-over-year growth rate from 2.1 percent to 1.8 percent. The trade balance is trickier to model, but those brave enough could estimate this by compiling published data from foreign trading partners.
Source: Saxo Bank
Source: Saxo Bank
Carlyle Group used proprietary data
The US-based asset manager Carlyle Group sits on plenty of data from its portfolio of more than 200 companies. The company published its own estimates of the missing data last Friday, and The Washington Post wrote about the release:
- Retail Sales: up 0.25 percent in September
- Consumer Price Index: up 1.5 percent YoY in September
- New Residential Construction: 913,000 housing starts annualised in September
- Durable Goods Orders: USD 68.6 billion in September
- GDP: up 1.7 percent in Q3 2013
The Fed will be in no hurry to taper
Even with a quick end to the shutdown, the data will not be deemed to be reliable enough for decision-making purposes. Probably at least a month has to pass after the shutdown has ended before the data will be reliable enough. Add in the fact that the hopefully temporary negative growth effects of the shutdown will take some time to drop off the data. Thus it becomes obvious that the Federal Reserve will not commence the tapering of its asset purchases any time soon. Note that if the government is opened and the debt ceiling is only raised temporarily, then uncertainties will continue for weeks to come and this would have a nasty negative effect on holiday sales, which are very important for the economy.
The earliest time for commencement of the Fed’s asset purchase tapering is now seen to be at the Federal Reserve Open Market Committee's January 28-29 meeting. That meeting is not followed by a press conference, so if the economy looks wobbly at the time there is a chance that the Fed would like to wait until its next meeting on March 18-19, after which it will hold a press conference. However, that date could be a bit too late for the Fed. Also note that there will be changes in the FOMC members: Charles Evans and Esther George rotate off the committee and Charles Plosser and Richard Fisher rotate on.
—Edited by Yvette Roper