When Will America Reopen

When Will America Reopen

The question was simple and on everyone's mind, yet impossible to answer with certainty. Any forecast was doomed to be partially, if not completely, wrong. The first vaccines were arriving, and new cases were still climbing. The Johns Hopkins chart of daily confirmed new cases on a seven-day average was horrific, with the line swimming off the chart. 1 became the definitive source for case counts and trends during the pandemic.

Before attempting to solve a problem, it helps to define what success looks like. Reopening was never a binary one or zero. It came in many shades of gray. How normal would normal be, with masks and social distancing or just like 2018? Where would reopening apply, in schools or restaurants, in urban or suburban or rural areas? How ubiquitous would adoption be, in pockets or everywhere? These questions framed the measurement challenge.

What Information Did We Have

When searching for Goldman Sachs and Covid-19, one resource that surfaced was a weekly newsletter called Measuring the Reopening of America. The newsletter tracked more than 100 metrics from 20-plus different sources. It presented trend data over four to five months, looking for acceleration or deceleration. It focused on leading indicators that offered hints about the future.

The newsletter went beyond macroeconomic and financial data. It augmented analysis with anecdotes from 14 co-authors from Goldman Sachs investment research. 2 detailed how the team sorted data into three big buckets and tracked unconventional metrics. The creativity lay in the structure and the breadth of the measurement.

Stay at Home: What We Started Doing

Think of all the new habits people started or increased over the preceding months. Working out at home, playing video games, baking bread and watching more streaming content all surged. Many people bunkered in their houses doing stay-at-home activities. The newsletter tracked grocery sales of refrigerated dough, cleaning supplies, staple goods and beverages. It measured online gaming through downloads, money spent and time spent. It followed app downloads for food delivery, at-home fitness, video chat and online payments.

These metrics quantified how insular daily life had become. The data told a story of adaptation, as households reorganized around the home. The strength of these signals indicated how far behavior had shifted from the pre-crisis baseline and how much momentum the stay-at-home economy carried.

Back to Normal: What We Stopped Doing

The flip side was equally telling. There was a litany of things people stopped doing. Gas demand dropped because there were few places to drive. Movies released to the box office showed a 100 percent decline during certain weeks. Restaurant reservations tracked through OpenTable collapsed. Cosmetics and shaving cream sales fell because nobody was commuting to an office. Hotel occupancy dropped from both business and leisure travel stopping.

These metrics measured the void left by paused activity. As they recovered, they would signal the return to normal. The pace of that recovery would vary by category, with some behaviors snapping back quickly and others lagging as caution persisted. Tracking the differential between categories revealed which parts of the economy were ready to reopen and which remained suppressed.

Business Activities

The US economy is enormous, exceeding 21 trillion dollars, so the list of potential metrics was understandably long. The newsletter tracked inputs to production like construction equipment. It monitored indicators of trade volumes such as trucks, rails and tires, including intermodal rail containers. It gauged business sentiment through metrics like advertising spend, which signaled whether companies felt bullish or bearish about consumer demand.

A few of the interesting metrics stood out. Domestic ethanol blending and Google search traffic provided easy-to-track signals. Home sales and inventory levels across both existing and new homes offered real estate insight. Pricing of trucking and the level of intermodal rail containers reflected supply chain activity. Dental volumes showed whether people were pushing out annual checkups. The ratio of branded versus generic prescriptions indicated whether consumers were cutting costs by buying generics.

What the Trends Look Like

The real power of data emerges over time. In the same way that a run chart is more useful than a pie chart, trends over four to five months revealed direction. Were there changes, or even acceleration and deceleration, in the metrics? Refrigerated dough sales alone were not a magical fortune-telling tarot card, but they served as a weak signal giving a sense of the trend. Added together with a hundred other metrics over 20-plus weeks, the combined picture told something meaningful.

The genius of the newsletter lay in three qualities. First, the breadth of measurement across 100-plus metrics. Second, the logical structuring of three buckets: home, normal and business. Third, the consistency of measurement across 21 weeks. 3 formalizes a similar approach for economic forecasting, combining multiple leading indicators into a single composite index. The Goldman Sachs approach applied that same logic to unconventional, high-frequency data.

Triangulate Answers Using Surrogate Data

These weak signals were easy to misinterpret individually. However, taken together longitudinally, they served like cell phone towers that let you triangulate a location. The triangulation produced a better set of answers than any single metric could offer. Any framework, tool, approach or scorecard can help make decisions, but it does not produce a multiple-choice answer. It generates insights that support good human judgment based on your values, risk tolerance and timeframe.

The methodology itself was the takeaway. Define what success looks like, identify diverse metrics across multiple categories, track them consistently over time and look for convergent signals. The approach transfers to any ambiguous question where no single dataset holds the answer. Consultants who master this technique deliver insight that clients cannot generate internally.

Dashboards Are Approximate

This approach was not perfect, and Goldman Sachs would not debate that point. Most metrics were tracked year-over-year, but after a year of Covid-19-shifted lifestyle, the baseline was dramatically different. Some metrics were tracked monthly or quarterly, which provided much less granularity. Markets and products get saturated, since you do not download a video conferencing app every week. Lots of interrelation exists between substitutes, new entrants and prices. Customer preference can permanently shift after extended periods, as seen in companies moving headquarters from California to Texas.

These limitations do not invalidate the approach. They simply mean the dashboard requires interpretation, not blind reading. The user must understand what each metric measures, what its baseline is and what distortions it carries. In conclusion, it was all about vaccines and smart living. Social distancing, mask wearing, hand washing and ventilation worked, and the path forward required staying smart and informed. Until then, the data kept flowing and the triangulation kept improving.

• 1Johns Hopkins Coronavirus Resource Center tracking data
• 2Goldman Sachs Research on measuring the reopening of America
• 3Composite leading indicators methodology from the OECD