Factual Friday!
Tariffs are down - right??; Bad data = Bad AI; WCRI webinar
Trump’s Trade War is…on pause? Over? Scaling back?
Conceding the Chinese aren’t going to bend over, this week Trump unilaterally slashed his Tariffs on Chinese goods from 145% to 30%.
Phew! Glad that’s over, eh?
Well, not so fast.
That 30% Tariff will apply to almost all consumer goods coming from China - and let us not forget Trump has imposed 10% tariffs on all other imports (except Russia and…North Korea !?)
While 30% and 10% don’t sound like much - they are a HUGE jump compared where we were before Trump started his Tariffs.
we’d expect Trump’s tariffs after last weekend’s retreat on China to cut overall U.S. trade by roughly 50 percent. Trade with China, which would have been virtually eliminated with a 145 percent tariff rate, would fall by “only” around 65 percent with a 30 percent tariff.
Does cutting U.S. trade with the world in general by half and reducing trade with China by two thirds sound to you like Trump calling off his trade war?
Fallout continues -
Walmart announced yesterday it is raising prices due to Trump’s Tariffs.
Trump threatened Barbies with a 100% Tariff after Mattel - which owns the Barbie brand - said it would raise prices due to Trump’s Tariffs.
Stanley - maker of insulated cups and power tools - said Trump’s Tariffs will force it to raise prices.
And baby stroller company UPPAbaby increased prices due to Trump’s Tariffs.
There’s relative change - 30% is a lot better than 145% - and absolute change - today’s Trump Tariffs are effectively way higher than the Tariffs that crashed the world economy before the Great Depression.
Takeaway - higher prices for consumers and business uncertainty bode ill.
If your data is bad, AI is pointless.
Excellent post from Stan Smith, CEO of GradientAI yesterday on WorkCompWire on a major issue in workers’ comp - data quality.
Stan notes:
Your AI is only as good as your data…An insurer my company works with recently found that 40% of its most critical data fields were missing or incomplete, making it difficult to build reliable AI models. This issue is widespread—missing or inconsistent data limits an organization’s ability to identify patterns and make informed decisions.
And I wholeheartedly agree; much of the work comp data we’ve worked with is pretty crappy…and it hasn’t improved over the last few years.
I’m afraid most in the industry will believe it doesn’t believe Stan’s assessment applies to them as their data is “fine”. The complacency trap is rampant in worker’s comp.
One other thought - many payers outsource data entry and management to third parties, parties that are selected in large part based on how cheaply they can intake claims/process bills/approve services/“manage” care. There are little to no efforts to assess, much less ensure vendor data quality.
But hey, combined ratios are great, so who cares?!
CompScope(TM) Webinar
WCRI’s hosting a webinar on Delaware’s workers’ comp system June 2 - no cost, and registration is here. Insider tip - this isn’t just about Delaware - there’s great insights into how the other 17 CompScope(TM) states compare to each other.