The case of the disappearing imports

Stephen Stanley | April 25, 2025

This document is intended for institutional investors and is not subject to all of the independence and disclosure standards applicable to debt research reports prepared for retail investors. This material does not constitute research.

When I was a kid, I loved reading mystery novels.  The first quarter GDP numbers bring me back to those days.  The monthly trade figures show clearly that firms were racing to get their merchandise imports into the country before tariffs were put in place.  Typically, when imports surge, from a GDP accounting perspective, they also show up in inventories, or possibly in consumer spending if they are bought quickly enough, creating an offsetting bump that cancels out the widening in the trade deficit.  However, in the first quarter, it looks like, barring a drastic inventory increase in March, the wave of imports did not find its way into inventories or consumer spending.  As a result, real GDP appears on a path to contract significantly in the first quarter, with the prospect of a sharp reversal in the spring. Beyond the spring, growth prospects look dangerously close to recession.

GDP accounting

The method that the Bureau of Economic Analysis uses for calculating Gross Domestic Product is complicated.  A decision was made decades ago that it was easier to compute and compile spending than output. Even though GDP measures domestic production, or output, the total is tallied up by collecting data on spending.  Note that the main components of GDP are consumer spending, business investment, and government outlays.

When consumer spending is calculated, no distinction is made with regard to whether each item bought was produced in the US or elsewhere.  It would be far too complicated for government surveys to ask Wal-Mart and other retailers to keep track of what proportion of their receipts came from domestically produced rather than imported goods.  As a result, it is more practical to tackle GDP from the spending side.

To get from the spending tally back to GDP, statisticians have to add in net exports. The equation comes right out of Econ 102: Introductory Macro: Y = C + I + G + X.  In essence, after the BEA tallies up the main elements of domestic spending, it needs to add exports and subtract imports to get to a proxy for US output.

Note that imports are, in theory, irrelevant to domestic production.  If a retailer imports a good, let’s say a pair of sneakers, from a foreign supplier, the direct impact on US production is zero, no matter what ultimately happens to those shoes.

However, given that GDP is tallied up from the spending side rather than the production side, when those sneakers are purchased from a shoe store, they add to consumer spending and thus to GDP.  To net that impact out, imports have to be subtracted.

Following the path of those sneakers chronologically, they are shipped into a US port.  At that point, they are recorded as an import, which means that they have a negative impact on GDP.  However, within days or weeks, the shoes will be transported to a warehouse, at which point they become included in wholesale inventories, or to a store, where they would be tallied in retail inventories.  Then, ultimately, they are bought, at which point they flow out of inventories and into consumption.  Throughout this process, there should always be a negative entry (imports) and a positive entry (inventories or consumer spending) that cancel out, so that the imported sneakers never meaningfully change the GDP number.  Exactly what that looks like in the quarterly GDP data may vary based on the fact that we measure GDP quarterly, making four (somewhat arbitrary) cutoff dates a year.  The later in the quarter that the import enters the country, the more likely that it is counted as inventory at the end of the calendar quarter, while the earlier, the more likely that the good actually gets bought and ends up in consumer or business spending.

Where did the imports go?

That discussion may seem tedious, but it was necessary to fully explain the odd developments in the first quarter.  As one might expect, firms were eager to rush their imports into the US in the first quarter to avoid tariffs that were ultimately imposed in April. It is not particularly surprising that merchandise imports exploded higher in the first few months of 2025.  Imports surged in January and remained at that inflated level in February.

There was an unusual complication in the import figures that I wrote about in detail recently.  A significant component of the bulge in monthly merchandise imports represented gold, as investors, fearing that gold would be subject to tariffs, wanted to move their physical gold holdings from overseas repositories, mostly in Switzerland, to the New York COMEX warehouses.  The reason this is a twist is that the BEA excludes the international flows in gold from the GDP calculations, reasoning that the bulk of gold exports and imports reflect investment decisions, not real activity.

The monthly trade figures through February, along with my projections for March, suggest that merchandise imports may have risen by about $550 billion on an annualized basis, an annualized rate of advance of between 65% and 70%.

Based on my estimates of the magnitude of gold imports, I am projecting that the GDP-consistent concept (imports excluding gold) may have risen by more like $220 billion.  That would still be an annualized pace of close to 40%.  In real terms, I look for the net exports component of GDP to subtract almost $200 billion from GDP in the first quarter, and I would not be surprised to see an even larger drag.

Ordinarily, as laid out above, there would be an offsetting rise in inventories or consumer and business spending that would offset the bulge in imports.  However, there is no evidence of that in the economic numbers so far.  Retail sales and core capital goods shipments through March did not point to a sharp run-up with the exception of auto sales, which did jump in March.

One would imagine that the bulk of the offset would come in wholesale and retail inventories.  To be sure, we do not have March data in hand, but the January and February numbers do not point to anywhere close to an offset to such a massive jump in imports.  The book value of factory, wholesale, and nonauto retail inventories rose by 0.5% in January and 0.2% in February.  I have penciled in an estimate of a 0.3% advance in March.

These monthly figures point to an acceleration in inventory accumulation after the last quarter of 2024’s unusually slow $9 billion annualized increase.  However, wholesale prices rose considerably in the first quarter, pointing to an inventory valuation adjustment that subtracts from the total.  In all, I look for an inventory rise of about $44 billion annualized in real terms in the first quarter.  That’s a decent-sized acceleration on the quarter but it comes nowhere close to fully offsetting the massive drag from net exports.

By my estimates, net exports may subtract over three percentage points from real GDP in the first quarter, while inventories could add just over half a percentage point.

Just to offer a sense of how sensitive these calculations are to my March estimates, I have assumed that wholesale inventories and nonauto retail inventories advance by 0.4% and 0.3%, respectively.  If both were instead to jump by 1%, then the impact on the quarterly inventory figure would likely be around $35 billion, worth another half a percentage point or so to first quarter real GDP, a meaningful difference, but, again, not even close to fully offsetting the imports spike.

First quarter GDP Implications

As a result of the mysterious disappearance of merchandise imports, presuming that most of the rise does not show up in either spending or inventories, I look for a noticeable decline in first quarter real GDP.  My current projection calls for an annualized drop for the quarter of 1.4%.

However, I also presume that merchandise imports are going to normalize in the spring, reversing the bulk of the widening in the real trade deficit in the first quarter.  As a result, real GDP may roar back in the second quarter, not because the economy is strong (on the contrary) but simply reflecting the swings in trade flows.  At the moment, I look for real GDP growth of better than 3% annualized in the second quarter.

Both of these results would be quite deceiving as it relates to the underlying health of the economy.  As I have often noted, the best gauge of the economy’s underlying pace is not real GDP but real domestic demand—GDP minus inventories and trade.  This gauge may post a tepid 1.3% annualized gain in the first quarter, less than half of the stellar 2024 pace of 3.0%.  From there, real domestic demand is likely to slow further, dragged down by the paralyzing uncertainty inflicted by the administration’s erratic tariffs policy.  I have penciled in annualized advances of only 0.2% in the second quarter and 0.7% in the third quarter, not quite a recessionary trend but dangerously close.