Goldman Sachs has reduced the probability of recession in the next 12 months to 30%, from 35%, and earlier 45% (Reuters<\/a>). It\u2019s important to note that these projections are conditional on\u00a0 the path of future policies \u2014 which in these times are less clear than ever.<\/p>\n Here\u2019s my math.<\/p>\n The current quota set for ICE deportations\/removals is 3000\/day. If that pace is achieved (which will almost assuredly scoop up more than just criminals), then after one year, about 1.1 million individuals will be removed.<\/p>\n McKibben, Hogan and Noland (2024)<\/a> provided simulation estimates for 1.3 million removed.<\/p>\n One interesting aside is that the impact on durable manufacturing employment seems rather large \u2014 much larger than that in agriculture, fishing and hunting. The idea that the impact would be largely limited to agricultural employment seems incorrect, in both % terms, and absolute numbers (there are about 800K workers in agriculture, about 8 million in durable manufacturing).<\/p>\n The 2026 impact looks rather modest \u2014 around 0.6 ppt deviation from baseline.\u00a0Note that the estimates are for deportations\/removals alone. They do not<\/em><\/strong> include tariffs. Assuming the current pause conditions hold, so universal 10% tariffs, with 45% or so on China (including past tariffs) hold, a lower bound estimate from trade policies is -0.3 ppts in 2026 \u2014 assuming no retaliation. With retaliation, its -0.9 ppts<\/a>.<\/p>\n One point is that none of these estimates incorporate the effect of policy uncertainty. However, we know that policy uncertainty is important \u2014 especially since policy uncertainty as measured by EPU is now about 4 times what is was pre-Trump. Assuming linearity, Ferrara and Guerin (2018)<\/a> estimates suggest unemployment will be about 2 ppt less than it otherwise would be.\u00a0 This implies roughly a 1.5 ppts lower GDP relative to counterfactual.<\/p>\n Add 0.6 to 0.9, and you get 1.5 ppts. But add policy uncertainty (which Trump seems to be trying to maximize), and that\u2019s 1.5 ppts added to 1.5 ppts to sum to 3 ppts. If baseline growth was estimated to be 2.3 ppts (I\u2019m using CBO January 2025 projection), then GDP growth going into 2026 will be negative.<\/p>\n So, I\u2019m doing a conditional forecast here: deportations\/removals hit close to 2.1 millions, tariffs are at least 10% universal, with retaliation, and policy uncertainty percolates through the economy as estimated using historical events. One point is policy uncertainty has not risen to these levels in recent times, aside from the once-a-century Covid Pandemic.<\/p>\n Figure 1:<\/strong> EPU (blue, left scale), VIX (red, right scale). NBER defined peak-to-trough recession dates shaded gray. Light orange denotes Trump administrations. Source: policyuncertainty.com, CBOE via FRED, NBER.<\/em><\/p>\n Figure 2:<\/strong> EPU-trade (blue, left scale), TPU (red, right scale).\u00a0 NBER defined peak-to-trough recession dates shaded gray. Light orange denotes Trump administrations. Source: policyuncertainty.com, Iacoviello et al.<\/a>, NBER.<\/em><\/p>\n Would I still rely on one-year-ahead recession forecasts from term spreads for the next six months? No, because the policy changes implemented in the past five months have been sufficiently unexpected enough (at least by the market), that the usual mechanism is unlikely to operate.<\/p>\n<\/p><\/div>\n![\"\"](\"https:\/\/econbrowser.com\/wp-content\/uploads\/2025\/06\/MHN2024fig1.png\")
<\/a><\/p>\n![\"\"](\"https:\/\/econbrowser.com\/wp-content\/uploads\/2025\/06\/MHN2024fig3.png\")
<\/a><\/p>\n![\"\"](\"https:\/\/econbrowser.com\/wp-content\/uploads\/2025\/06\/MHN2024fig6.png\")
<\/a><\/p>\n![\"\"](\"https:\/\/econbrowser.com\/wp-content\/uploads\/2025\/06\/epu_vix.png\")
<\/a><\/p>\n![\"\"](\"https:\/\/econbrowser.com\/wp-content\/uploads\/2025\/06\/eputrade_tpu.png\")
<\/a><\/p>\n