With an “A/B testing” president in office, financial models may offer clarity
It’s a new era for business as this week marked the beginning of an all-out trade war between the US and Canada. Emotions were high, and crisis-tested leaders were in fashion as a resurgent Liberal Party in Canada appointed central banking legend Mark Carney prime minister. Enter a macroeconomic technocrat who weathered the Great Recession and Brexit, battles many hope have prepared him for his next act: navigating a trade conflict with an ally whose leader has threatened to annex Canadian territory.
Barb-trading amongst national leaders has continued, with a “tariff on/tariff off” jockeying for position that’s resulted in the rates doubling and halving depending on the hour (see chart below). As of Wednesday, both Canada and the EU had announced reciprocal tariffs on steel, aluminum, and a variety of American-made goods.
Domestically, expectations of a slowing economy have transitioned to the specter of recession, with top analysts on Wall Street breaking out the “R” word this week in force as market turbulence continued. President Trump appears to agree, having refused to rule out the possibility that tariffs could squash his economy. Some experts see this as a grand policy experiment — but allow that it could trigger a dangerous chain reaction.
“I often call Trump … the president of A/B testing: put stuff out into the world, see what reacts, iterate on it, you know, get some feedback, and put something more official out, and so on,” Leif Abraham, Co-Founder and Co-CEO of trading platform Public, said Tuesday onstage at Fintech Meetup. “Markets obviously don’t like that, right? In markets, you want as clear a path as possible due to forecasts.”
A/B testing can be profoundly costly when a $30 trillion economy is at stake — it’s not exactly a minimum viable product. In an email interview with Fintech Nexus earlier this month, Ernie Tedeschi, Director of Economics of The Budget Lab at Yale and former Chief Economist at the White House Council of Economic Advisers, said that the unpredictability of tariffs’ implementation may counterbalance any consolatory benefits that come from their having been forewarned.
“A longer lead time could in theory mitigate the effect of the tariff somewhat because it gives firms and consumers time to make purchases & plans … So the one-month ‘pause’ of the 25% Canada/Mexico tariffs may have blunted their short-term pain a bit,” Tedeschi told Fintech Nexus. “But the uncertainty around trade policy has been so high it’s unclear that helps much these days: Markets and consumers aren’t sure what tariff threats will actually stick, so it’s very possible they didn’t take the Canada/Mexico tariff threat seriously.”
The Budget Lab recently estimated that a “reciprocal” tariff policy would lead to per-household consumer losses between $2,700 and $3,400 in 2024, would shrink the US economy by 0.3% and 0.6% in the long run, and would result in lost tax revenue somewhere between $200 billion and $400 billion. Tariff-induced contractions would also affect lower-income households three times as severely as it would the top 10% of households by income. Economics being its own kind of experimental science, much of what we think we know about who and what gets hit from tariffs and by how much depends on what metrics are added to the models. And a large shock — like pandemic lockdowns or blanket tariffs — can upend predictive qualities.
For example, Tedeschi said the Lab’s worst-case estimates assume “Canadian (and Chinese and Mexican) domestic politics are such that their governments choose a full tit-for-tat response, despite the fact that such a response actually makes the trade war more painful for everyone including them.” (The Lab looked at European economic responses in its report as well.) The models use historical evidence from each country to quantify the outcomes of campaigns like “Buy Canadian” as well as reshoring effects.
These models — financial technologies in their own right — used data from prior trade wars, and the conclusion is things are only going to get uglier if the US government sticks to its current path. It begs the question; has the US economy ever bounced back from a major slide as quickly as it’s fallen? Using a different model as a stand-in for economic growth suggests that’s pretty unlikely.
That model, GDPNow, is an unofficial weekly forecast of real GDP growth from the Atlanta Fed, and has fallen more than five percentage points over the past month, going from a 2.9 percent estimated growth rate on February 13 to a -2.4 percent predicted decline in GDP as of March 6.
Fintech Nexus asked Patrick Higgins, a policy adviser and economist at the Atlanta Fed who helped develop GDPNow, whether these estimates have ever recovered quickly from quantitative declines to return to their previous trending line. (Note: We interviewed Higgins via email before Trump’s recession-flirty statements and before the Federal Open Market Committee’s blackout period began.)
“I would be surprised if there has been this sharp a move that has been quickly reversed,” Higgins said. GDPNow previously fell three percentage points between June 28 and July 1 in 2022, and, according to Higgins, “ultimately reversed some, but not all of that drop.” GDPNow’s downward turn at the end of February was largely “concentrated in net exports,” whose contribution to net growth fell about 3.3 percentage points, Higgins added.
Higgins said in 2022 that systemic shocks like the Covid-19 pandemic have impacted the accuracy of GDPNow. Gauging its utility today, it’s important to note that GDPNow does not have potential policy-induced macroeconomic shocks built into its model, which has stayed consistent since 2023. It doesn’t recognize a major jolt like fiscal stimulus until it shows up in monthly spending data. “Assuming no data releases, the model forecast would be the same just after a policy change than it would be before,” Higgins told Fintech Nexus. GDPNow also leaves out more qualitative data like consumer sentiment or uncertainty.
A subset of fintechs whose businesses benefit from Rothschild-esque “the time to buy is when there’s blood in the streets” market dynamics might see a looming “R” word as the dawn of a bespoke bull market.
“Betterment really thrives in turbulent markets in large part because tax loss harvesting is something we do best, and the value creation really jumps off of volatility,” said Sarah Kirshbaum Levy, CEO of wealth-management platform Betterment, onstage Tuesday at Fintech Nexus.
A bullish-to-bear-markets clique might emerge prominently if a recession becomes reality — corporate vessels exempt from a lowering tide’s negative implications. Dori Yona’s SimpleClosure would probably do well with a wave of failed ventures looking to fully shutter. Earned-wage access platforms might see surges in use, as might alternative funding services for companies kicking the can on an IPO. That doesn’t necessarily bode well for the everyday consumer, business, or national economy. Maybe their performance should become a reversely correlated index or model for economic health.
