There is still a huge amount of uncertainty about the future of the outbreak and the economic impacts associated with it, but we do have the first hard data points. The forecast assumes that the worst of the outbreak is over and that there will be no major resurgence in the fall.
Overall, the forecasts for Arizona, Phoenix, and Tucson have been marked up from the April projections but the outlook still calls for a major downturn. The magnitude is now expected to be similar to that of the 2008-2009 recession.
The recovery is forecast to begin later this year. The level of economic activity is expected to return to pre-outbreak levels by early 2022. I characterize this as a “Nike-swoosh” type recovery.
Arizona Recent Developments
Arizona nonfarm payroll jobs experienced a huge decline of 232,000 jobs over the year in April, according to the preliminary estimate. That translated into a decline of 7.9%, which was less than the national decline of 12.9%.
Arizona jobs declined by 288,000 over the month in April, using BLS seasonally-adjusted data. From February 2020, which appears to be the peak in the seasonally-adjusted data, Arizona has now lost 310,000 jobs.
Arizona’s seasonally-adjusted unemployment rate (U3) spiked up significantly in April, reaching 12.6%. That was below the national rate of 14.7%, but was the highest reading since the beginning of the current series in 1976.
Across Arizona’s metropolitan areas, Flagstaff experienced the largest job decline in April, while Sierra Vista-Douglas registered the smallest. Flagstaff leisure and hospitality employment declined by 59.1% over the year, accounting for 75% of the losses in the metropolitan area.
Nonfarm payroll jobs declined by 7.6% over the year in the Phoenix MSA in April. The Tucson MSA experienced an 8.6% decline during the same period.
Leisure and hospitality lost the most jobs over the year, dropping 126,000 jobs. Employment at food services and drinking places decline the most. Trade, transportation, and utilities lost just over 41,000 jobs over the year, with widespread losses in retail (except for food and beverage, building supply stores, and other general merchandise). Professional and business services lost over 25,000 jobs, with administrative and waste services accounting for most of the decline. Education and health care lost 19,000 jobs over the year, with losses concentrated in health care and social assistance.
Keep in mind that monthly employment releases always contain revised data for the prior month. With the April release, the U.S. Bureau of Labor Statistics revised down its March nonfarm payroll estimate for Arizona. The nonseasonally-adjusted data now show a 20,700 job decline in March, compared to a 7,400 job decline in the preliminary data. Arizona’s March unemployment rate estimate was revised up to 6.1% from 5.5%. During downturns, it is not unusual for preliminary estimates to be revised down in the next monthly release.
For the current Arizona forecasts, the baseline projections are assigned a 45% probability. The pessimistic scenario has a 35% probability and the optimistic scenario has a 20% probability.
A national forecast from IHS Markit, summarized below, is combined with my assumptions about the impact of social distancing on employment and retail sales. These assumptions are scaled to reflect the latest information available.
Based on the April employment release for the state, I have significantly revised the estimated impacts. For leisure and hospitality and retail trade, the impact is less severe than expected in April. The drop is still huge, just not as bad as expected in the April projections.
Taxable sales at restaurants and bars are forecast to experience a huge decline in the second quarter. Retail sales also decline in the second quarter, but the drop is much smaller than projected in April. This upward revision reflects better than expected performance in March/April.
The current baseline projections put the magnitude of the coming downturn at about the same size as the 2008-2009 recession but much shorter in duration. Under current assumptions, it is a short, sharp shock. Total nonfarm employment in Arizona declines by 386,000 from the first quarter to the second quarter of 2020. That translates into a peak-to-trough decline of 11.3%. For comparison, Arizona lost 305,000 jobs from peak to trough (quarterly) during the Great Recession. That was an 11.4% drop.
The state unemployment rate surges to 17.0% by the third quarter of 2020, but then begins a gradual decline.
Taxable retail sales decline this year, as job, income, and wealth declines (and social distancing) take a toll on consumers. Sales at restaurants and bars gets hit harder, given the more severe headwind from social distancing.
Population growth slows, with reduced net migration, and that generates lower levels of housing permit activity.
Once we get past the shock this year, the recovery should be steady. Arizona was in very good shape before the outbreak and once the outbreak is under control, growth will accelerate significantly. However, it will take time to get back to pre-outbreak levels of economic activity.
Growth rates will look very good initially, because we’re restarting large sectors of the economy. The forecast calls for the level of employment to return to its first quarter 2020 level by the first quarter of 2022. I characterize this as a “Nike-swoosh” recovery.
With the high degree of uncertainty surrounding the outlook, it is a good time to pay attention to the alternative scenarios.
The pessimistic scenario assumes a deeper, more prolonged downturn. Arizona jobs decline by 428,000 from the first quarter of 2020 to the second quarter of 2020. That translates into a 12.7% drop.
The state unemployment rate peaks at 23.3% in the fourth quarter of 2020, then gradually declines.
Under the pessimistic scenario, it takes longer for the state to return to pre-outbreak levels. Jobs return to their first quarter of 2020 level in the first quarter of 2024, four years later.
The optimistic scenario calls for state nonfarm payroll jobs to decline by 346,000 from the first to the second quarter of 2020. That is a 9.9% decline. The unemployment rate rises to a peak of 15.4% in the third quarter.
Employment recovers faster under this scenario, with jobs reaching their prior peak by the second quarter of 2021. That would qualify as a “V” type recovery, in my view.
The May projections from IHS Markit assume that the national number of COVID-19 cases and deaths fall to levels low enough to permit fairly wide relaxation of social distancing by June. This is a month or two earlier than presumed in prior projections.
Even with that somewhat more optimistic assumption, the baseline forecast has been revised down again, particularly for the labor market indicators.
The forecast now calls for year-over-year employment to fall by 13.5% in 2020, compared to 4.5% in the April forecast. The unemployment rate is now forecast to peak at 19.6% in the third quarter of 2020.
Real GDP growth is expected to drop by 36.5% in the second quarter, at a seasonally adjusted annual rate. That is the rate at which real GDP would decline if the quarterly rate were sustained for an entire year. It is driven by an even larger drop in real personal consumption expenditures, which fall by 42.6% in the second quarter, annualized.
On a year-over-year basis, real GDP is forecast to decline by 7.3% in 2020, the bounce back to 5.1% in 2021 and 4.6% in 2022. It takes until the second quarter of 2022 for real GDP to return to the level last seen in the final quarter of 2019. IHS Markit now characterizes the recovery as “U” shaped, which refers to the evolution of the level of real GDP during the forecast.
Real personal consumption expenditures plummet by 8.3% this year, driven by at-risk categories related to travel and tourism. From the fourth quarter of 2019 to the second quarter of 2020, real spending on accommodation services falls by 60.1%. Real spending on food services falls by 45.3%. Real recreation spending is forecast to decline by 44.0%, and real public transportation spending falls by 66.1%.
Need more information? Contact Dr. George Hammond about joining the Forecasting Project.