Look around and see enough metrics (like vehicle mileage) You might think so. Let’s take a look at what regular and high frequency indicators mean.
Here are the key metrics tracked by the NBER Business Cycle Dating Committee, as follows:
figure 1: Nonfarm payrolls (dark blue), as of June 25, Bloomberg consensus on nonfarm payrolls (ble+), industrial production (red), mid-2012 excluding transferred personal income (green), mid-2012 Manufacturing and trade sales $ (black), February 2012 consumption (light blue), and 2012 monthly GDP (pink), all log normalized to 2020M02=0. NBER defines recession dates, peaks and valleys, shades of gray. Source: BLS, Federal Reserve, BEA, via FRED, IHS Markit (nee Macroeconomic Advisers) (published June 1, 2022), NBER, and author’s calculations.
What about indicators of weekly frequency? This is the Lewis-Mertens-Stock metric.
resource: Fredaccessed 25 June 2022.
Here’s a description of the index:
WEI is an index of real economic activity using timely and relevant high-frequency data. It represents the common components of ten different daily and weekly series covering consumer behavior, labor market and production. WEI is proportional to fourth-quarter GDP growth; for example, if the WEI reads -2%, and the current WEI level persists for the entire quarter, one would expect GDP in that quarter to be, on average, 2% lower than a year ago.
WEI is a composite of 10 weekly economic indicators: Redbook same-store sales, Rasmussen Consumer Index, new unemployment insurance claims, ongoing unemployment insurance claims, adjusted income/employment tax withholding (from Booth Financial Consulting), Rail Transportation (from Association of American Railroads), American Staffing Association Staffing Index, Steel Production, Wholesale Sales of Gasoline, Diesel, and Jet Fuel, and U.S. Weekly Average Electricity Loads (remaining data provided by Haver Analytics). All series are expressed as year-over-year percent change. These series are combined into a single weekly index of economic activity.
For more details, including an analysis of model performance, see Lewis, Mertens, and Stock (2020), “US Economic Activity during the Early Weeks of SARS-Cov-2 Outbreak”.
Here’s the OECD Weekly Tracker for the US:
resource: OECDaccessed 25 June 2022.
Partial description:
Nowcasting with Google Trends
Use machine learning to extract and aggregate signals about multiple aspects of the economy from Google Trends to infer a timely macroeconomic picture. The algorithm extracts and compiles information about consumption (e.g. from searches for “vehicles”, “home appliances”), labor market (e.g. “unemployment benefits”), housing (e.g. “real estate agency”, “mortgage”), business services (e.g. “Venture Capital”, “Bankruptcy”), industrial activity (e.g. “maritime”, “agricultural equipment”), trade (e.g. “export”, “freight”), and economic sentiment (e.g. “recession”) and poverty (e.g. “food” bank”). Using a number of variables can reduce the risk associated with structural breakage in a particular series, as highlighted by the failure of the “Google Flu” experiment.
Weekly Tracker uses a two-step model to forecast weekly GDP growth based on Google Trends. First, the quarterly model of GDP growth is estimated by quarterly search intensity based on Google Trends. Second, the relationship between Google Trends and activity, using the same elasticity estimated from the quarterly model, was applied to the weekly Google Trends series to produce a weekly track. Therefore, the OECD Weekly Tracker can be interpreted as an estimate of the year-on-year growth rate of “weekly GDP” (the same week compared to the previous year).
What about other big data indices? Here is the Brave-Butters-Kelley Big Data (500 Time Series) Concurrent Index as of April (May index released at the end of the month):
resource: Fredaccessed 25 June 2022.
The index is described like this:
The Brave-Butters-Kelley Index (BBKI) is a research project of the Federal Reserve Bank of Chicago. BBK US’s Concurrent and Leading Index and Monthly GDP Growth is constructed from a dynamic factor analysis of a set of 500 monthly indicators of real economic activity and quarterly real GDP growth.
The Backgammon Concurrent Index is the sum of the leading and trailing subcomponents of the cycle, measured in units of standard deviation of the real GDP growth trend.
As of April, the index was above trend, so there was no negative growth.
It would be nice to have more recent big data observations. Here are Google’s weekly liquidity metrics:
resource: Slok, Agarwal, Shah, deceleration watchApollo (Jun 22, 2022), Offline.
As to whether a slowdown in VMT portends a recession (eg gentlemen.Kopitz), well, it’s hard to tell. This picture suggests that gasoline prices may be related to a decline in VMT (remember, inflation is not the same as recession).
figure 2: Year-over-year growth in vehicle miles traveled (VMT) (blue, left scale) and real gasoline prices (tan, right log scale). Gasoline prices are affected by core CPI. The NBER uses shades of grey to define the peak and trough dates of the recession. Source: DoT, EIA, BLS by FRED, NBER and author’s calculations.
I estimate the short-run price elasticity of the VMT to be about -0.19 (using the first log-difference norm for the 2000-22 sample period, including consumption as the activity variable and covid variable, with a price coefficient at the 95% level).
All of this data shows me that we are not in a recession as of May. And as of June, we’re also unlikely to be in a recession, although the fact that data revisions are a reality makes me a little cautious. Are we going to be in a recession next year, well, the indicators there are more worrisome.
