Well he doesn't know how to divide. If private sector does 2% and government does 1% then government is 50% not 33%. Similarly he displays the level of government spending, not the change. If government is 30% of GDP, then automatically its 30% of GDP growth no matter what. And again 30% is really 50% of the private sector. Now if the government grows from 30% to 35% like it shows on that chart, in addition to few quarters of 50%, thats how you get to government being 100% of the growth. Just use $, the % is clearly too hard for HN.
Economy was 100. Government 30, private sector 70. Economy grew to 105, government 35, private sector 70. So 100% of the growth was government.
I hope the author reads this comment, HN is being a weird echo chamber. Bayesian probability people, what is the chance that All In is actually wrong based on a random blog post.