Outdoor Power Equipment Institute Annual Meeting
Peter Rupert delivered an economic update at the OPEI conference in Savannah, Georgia on June 25, 2026. He opened by acknowledging the obvious: there's a lot of uncertainty right now, between war, tariffs, taxes, inflation, and debt.
But he was quick to note that uncertainty is the constant, not the exception. His recommendations were refreshingly simple: skip the evening news, stay calm, and look for reasons to be optimistic.
To make the point concrete, he showed just how frequent recessions have been across U.S. history, and set that against the long arc of real GDP growth, which has climbed steadily for over 150 years despite a dozen of downturns along the way.
Using a mix of payroll employment, industrial production, personal income, and manufacturing and trade sales, Peter walked through a model estimating the probability the economy is currently in a recession. The reading is low, well below the spikes seen in 2001, 2008, and 2020.
On trade, Peter's framing was straightforward: tariffs are a tax on imports, and the costs get split between producers and consumers. He raised pointed questions about where the current tariff figures come from, and used the 2018 steel tariffs as a case study. Employment in iron and steel has been on a long downward slide since the late 1980s, and the tariffs don't appear to have reversed that trend in any lasting way, even as output per worker in the industry has fallen sharply since 2018.
Peter then switched gears to the labor market and showed a full year of monthly payroll forecasts next to what actually happened, and the record is rough. Forecasters missed nine times out of twelve, often by wide margins (February 2026 was expected to show a gain of 50,000 jobs; the actual number was a loss of 156,000). He offered a few explanations, a 2025 government shutdown that disrupted BLS data collection, sizable benchmark revisions, and a "low-hire, low-fire" labor market that's just hard to model, but his honest summary was that nobody really knows. Despite the volatility, he noted the labor market has shown a bit of a rebound since January, and that job openings and unemployed persons are currently roughly balanced, one job vacancy per unemployed person.
There are three potential risks on the horizon that Peter flagged: inflation, debt and interest payments, and AI. On inflation, he pointed to CPI and PCE both ticking up in 2026 after a relatively calm 2025. A continued rise would likely mean higher interest rates, especially if GDP and the labor market keep strengthening. On debt, he showed federal debt held by the public climbing back toward 100 percent of GDP, with interest payments, transfer programs, and defense spending as the major drivers.
The back half of the talk was devoted to AI and jobs. He walked through every major technological revolution since 1800, steam and rail, electricity, the automobile, agricultural mechanization, and computers, cataloguing the jobs each one destroyed and the jobs it created. Handloom weavers and stagecoach drivers gave way to factory and rail jobs. Gas lamp lighters and ice deliverymen gave way to electricians and appliance technicians. Farm laborers, once ~75% of the workforce in 1800, now make up about 1.5%, yet the unemployment rate today is roughly what it was in 1800.
His favorite illustration was the Great Horse Manure Crisis of 1894, when experts confidently projected that Manhattan would be buried under nine feet of manure by 1930. The problem was visible; the solution, the automobile, was completely unimaginable at the time. He argued that's roughly where things stand with AI today.
He also pushed back on the "this time is different" instinct, pointing to Keynes warning about "technological unemployment" in 1930, Leontief predicting in 1983 that computers would do to office workers what tractors did to horses, and Geoffrey Hinton declaring in 2016 that radiologists should stop being trained because deep learning would surpass them within five years. A decade later, radiologist employment and demand are both at record highs, since reading images turned out to be one task among many that make up the job, and AI tools have augmented radiologists rather than replaced them.
He didn't dismiss the costs. Transitions are painful and unevenly distributed, and some communities might suffer for a generation or longer. But on the aggregate numbers, the case is overwhelmingly positive: real GDP per capita is up roughly 30 times since 1800, life expectancy has roughly doubled, annual hours worked have fallen from about 3,000 to about 1,800, and food as a share of household budget has dropped from over 40% to under 10%. As he put it, nobody would actually trade their current standard of living for that of the median American in 1925, even at the same relative income percentile.
AI will cause real restructuring. Some occupations will shrink or disappear, and new ones will emerge, most of which don't have names yet. The right policy response, in Peter's view, is to cushion the transition rather than try to halt the technology. As he put it, the historical record on this question is unanimous, and it's on the optimistic side.