We need to talk about the twin deficits

Recent federal deficit and tariff revenues

The US Treasury recently released its monthly budget statement. While not as exciting as the events disclosed in my secret diaries, this revealed little progress in reducing the deficit despite tariff revenue receipts.

The July Federal deficit was $291bn, worse than expected and wider than the same month last year. In the 12 months to July the deficit is 6.4% of GDP, only a slight improvement from the 7% of GDP at the start of the year.

The deficit now includes tariff revenues which in July were running at around 1.1% of GDP (annualised). With goods imports at 11% of GDP in the second quarter of 2025, this implies a current effective tariff rate of 10%. However, the static calculated effective tariff – based on policy announcements – is around 15%. 

There is also the prospect of potential sectoral tariffs on pharmaceutical and semiconductors. While there might be some further substitution away from tariffed goods, we expect tariff revenue to rise to nearly 1.5% of GDP later this year.

The puzzle of who is paying for the tariffs

We have high confidence in the tariff revenue data. Someone has to be paying for them. So far the consumer has been relatively unscathed as we see only two or three tenths in extra inflation from tariffs. However, there is not much evidence of import prices being cut (required if foreigners are to 'eat the tariffs'). 

This leaves corporate margins as the remaining candidate. Whole economy (NIPA) profits are published with a lag. The most recent data only covers as far as the end of the first quarter of 2025, so does not yet capture the effect of the changes in tariffs since. Corporate tax revenue is down 6% in the fiscal year so far (to the end of July) versus the same period last year, but yet S&P500 companies are reporting strong earnings (although this is heavily skewed to big tech). 

Perhaps it is the unlisted sector getting hit? This could potentially lead to weaker economy wide hiring and investment (outside of the big tech AI investment surge). We also expect more of the tariffs to be passed through and a bigger increase in inflation later in the year.

Long-run implications

The budget deficit is unlikely to significantly improve as the front-loaded deficit spending (worth almost 1% of GDP in cost, but not stimulus due to a low multiplier) in the One Big Beautiful Bill takes effect over the next few months and into 2026. 

Some further increase in tariff revenues and higher capital gains taxes (if buoyant equity markets can be sustained) could help keep the deficit relatively steady over the next year. The slight improvement beyond 2029 shown in the chart below relies on the unrealistic expectation that the new tax cuts scheduled to expire are not extended.

The tariffs have become essential revenue raisers. As such we believe that any future administration will find it difficult to reverse tariffs without either raising taxes or cutting spending. Furthermore, the effectiveness of tariffs in reducing the trade deficit is blunted if the budget deficit remains around 6% of GDP as this tends to pull in imports – hence the twin deficits. 

With a large budget deficit, the national accounting identity means reducing the trade deficit would require an increase in private sector saving. This would likely to lead to weaker domestic demand – though stronger exports could be the preferred remedy.

Investment commitments into the US (as part of the trade 'deals') also in theory make it harder to reduce the trade deficit. The financial account has the equal and opposite sign to the current account. In practice, the investment commitments are gross, not net flows, often over unspecified time horizons and in some cases could end up like the Chinese agreements to buy US products in 2019  that never materialised.

Overall, and despite the policy changes and stated goals, we don't expect much progress on tackling the US twin deficits over the next year (outside of reversing the sudden widening in the current account deficit in the first quarter of 2025 due to the front loading of imports ahead of tariffs).

This is not a long-run sustainable equilibrium, in our view, with multiple paths before something potentially breaks. Meanwhile, markets appear to be expecting an AI-generated productivity uplift to prove the salvation.

Assumptions, opinions, and estimates are provided for illustrative purposes only. There is no guarantee that any forecasts made will come to pass.