Freight and Packaging Indicators
Table of Contents
Freight and packaging indicators can reveal pressure in the goods economy before it becomes obvious in a company’s earnings. Fewer shipments can matter to a trucking business, while weaker container production can matter to a packaging supplier. But a falling freight bill, a falling shipment count and a falling paperboard-production index describe different developments. Reading them as the same recession signal can lead to the wrong conclusion.
This guide compares the construction of several widely available U.S. and North American measures. The aim is to identify what changed, whose activity was measured and how far the evidence can support an economic or investment decision.
What Do Freight and Packaging Indicators Measure?
A shipment count measures movements, transportation expenditures measure dollars spent, a freight price index tracks specified service prices, and packaging output measures production. None is a direct measure of all final goods and services produced in the economy.
There is no single universal “freight index” or official economy-wide “cardboard box index.” A reference to either needs a provider, series name, unit, period and adjustment basis. Even physical measures differ: ten small shipments need not contain more goods than one large shipment; tons measure weight, and ton-miles also incorporate distance.
That distinction matters when comparing these series with gross domestic product. Goods can be transported between several businesses before a final sale, while services can add economic output with relatively little associated freight. Freight activity is useful evidence about a part of the economy, rather than a quantity to add directly to GDP.
Shipment Volumes vs Expenditures vs Freight Prices
The following measures answer different questions. An index expresses change relative to its own reference period; its level is not a dollar price or a directly comparable quantity across providers.
| Measure | Construction and units | Main interpretation limit |
|---|---|---|
| Cass Freight Index: Shipments | Index of shipment counts in Cass client transactions; January 1990 = 1.00. | A client-based count, rather than tons, ton-miles or every North American freight movement. |
| Cass Freight Index: Expenditures | Index of nominal transportation spending in those transactions; January 1990 = 1.00. | Both the amount of activity and the cost and composition of service affect spending. |
| BTS Freight Transportation Services Index | Seasonally adjusted index combining modal physical activity using value-added-based weights and chaining; 2000 = 100. | Specified for-hire modes; neither a shipment-count total nor a complete measure of freight. |
| BLS industry PPI: General freight trucking, long-distance truckload | Sampled selling-price index; December 2003 = 100; monthly, not seasonally adjusted. | A specified trucking industry, rather than all freight services or freight volume. |
| Federal Reserve paperboard-container industrial production | Real-output index, 2017 = 100, using physical shipments with an inventory adjustment. | An estimated manufacturing output measure, rather than the number of filled boxes sold to consumers. |
Definitions and construction: Cass, Bureau of Transportation Statistics (BTS), Bureau of Labor Statistics (BLS), and Federal Reserve documentation. [1] [2] [5] [7]
Cass: client shipments and spending
Cass covers for-hire, intra-continental North American freight across domestic modes, using transactions from its freight-payment client base. Its spending measure includes contract and spot business. Observations belong to the month Cass processes the invoice, which can differ from the shipment month. Cass adjusts for a 21-day work month and changes in its business base. These are useful construction features, but they do not make the observations a census of the economy. [1]
Timing and weighting help explain why Cass can disagree with another freight measure. A processing delay can move recorded activity between months. A count also gives a small shipment one movement just as it does a large one; its economic meaning differs from a measure based on weight or distance.
BTS: a weighted physical-output measure
The Freight Transportation Services Index, or Freight TSI, combines trucking, freight rail, inland waterways, pipelines and air freight. BTS seasonally adjusts the modal inputs and combines them using weights derived from the modes’ value added, with Fisher chaining. Its objective is physical transportation output; where ideal ton-mile data are unavailable, the inputs use proxies such as tonnage. Its trucking input, for example, comes from an American Trucking Associations member survey. [2]
The coverage excludes private trucking, parcel/courier services, the Postal Service and international or coastal steamship movements. A retailer shifting deliveries between its own fleet and a covered for-hire carrier can therefore change the relationship between the index and its total distribution activity. More comprehensive aggregation still has a boundary. [3]
BLS: prices for specified services
A freight Producer Price Index addresses a different question: how have the selling prices received by producers of the specified service changed? BLS selects establishments and transactions through probability sampling and weights observations. The specification of the service, buyer and transaction terms helps distinguish price change from simply selling a different mix of services. [5] [6]
For a concrete series, the BLS industry PPI for general freight trucking, long-distance truckload is PCU484121484121, published monthly without seasonal adjustment on a December 2003 = 100 base. It is narrower than all trucking or all freight. [16]
A price index is neither an average dollar quote nor a carrier profit margin. Dividing broad multimodal spending by one narrow trucking PPI does not automatically produce a valid quantity measure: the services, weights and charge coverage must match. This is the same need for a suitable deflator that underlies real versus nominal GDP, applied to a narrower service basket.
A Worked Example: Why Spending Per Shipment Can Mislead
First consider a hypothetical shipper over two equally long periods. Every movement is a comparable service: routes, distance, weight, mode, customer coverage and contract mix stay fixed. Fuel and extra service charges are included on the same basis. Only the number of movements and the uniform all-in charge change.
Period 1: 100 shipments × $200 = $20,000.
Period 2: 90 shipments × $240 = $21,600.
Shipment count changes by (90 / 100 − 1) × 100 = −10%, while the comparable charge rises (240 / 200 − 1) × 100 = 20%. Spending rises (21,600 / 20,000 − 1) × 100 = 8%: 0.90 × 1.20 = 1.08.
Indexing both count and spending to 100 in period 1 gives period-2 values of 90 and 108. The growth in spending per shipment can be recovered from their relative changes:
Under the fixed-composition assumptions, that result matches the charge increase. With separately normalized indexes, the ratio can show relative movement; it does not reveal the actual dollar charge without the underlying spending and count information.
Now change the mix while holding service prices fixed
Reset the example. There are two services: a short route priced at $100 per shipment and a long route priced at $300. Both prices remain unchanged. Each period still has 100 shipments, but the share on the short route rises.
| Quantity or amount | Period 1 | Period 2 |
|---|---|---|
| Short-route shipments at $100 | 50 | 80 |
| Long-route shipments at $300 | 50 | 20 |
| Total shipments | 100 | 100 |
| Total spending | $20,000 | $14,000 |
| Average spending per shipment | $200 | $140 |
Period 1 costs 50 × $100 + 50 × $300 = $20,000. Period 2 costs 80 × $100 + 20 × $300 = $14,000. Dividing each total by 100 shipments gives $200 and $140. Average spending falls (140 / 200 − 1) × 100 = −30%, although neither specified price falls.
Cass calculates inferred rates as its expenditure index divided by its shipment index. Its report identifies fuel, modal/within-mode mix and accessorial charges as influences. Interpret the result as relative cost per shipment; the route example shows why that need not equal a matched-service price change. [13]
What Does a Cardboard or Packaging Indicator Represent?
Start by identifying the production stage. Containerboard is material used in corrugated packaging; a container manufacturer converts material into products. Mill tonnage, container square footage, shipments, inventory and production capacity are related observations, with different units and timing.
The Federal Reserve’s paperboard-container output measure illustrates why a familiar label needs methodological detail. Its source table identifies square-foot shipments of corrugated and solid-fiber products, with a model-based inventory adjustment, using Fibre Box Association and Federal Reserve information. Paperboard-mill output instead uses tonnage from AF&PA. Thus an estimated container-production index is not a literal count of boxes, and the mill and container series measure different stages. [7]
For the American Forest & Paper Association (AF&PA), distinguish an annual capacity survey from production releases. Its 2024 capacity report describes mill coverage representing about 89% of U.S. paper and paperboard capacity, with estimates completing the dataset. That is the stated coverage of that survey, not proof that every separate monthly packaging release has the same sampling fraction. Total paper and paperboard also includes nonpackaging grades. [10]
Production growth need not equal domestic demand growth
AF&PA’s April 26, 2024 release provides a concrete example. First-quarter containerboard production rose 7% from Q1 2023, while domestic new supply rose 3.7%. The report also described a sharp increase in production of unbleached kraft liner for export. Different trade flows and measures can therefore produce different growth rates; the headline production increase was not a measured 7% increase in U.S. final consumer demand. [9]
Plant openings, closures and downtime also affect the relationship between capacity and actual production. Read output alongside the relevant capacity-utilization measure rather than interpreting an operating-rate change as if it were an identical change in tonnage. The underlying inventory-adjustment mechanism belongs in our guide to the bullwhip effect and working capital: destocking can reduce replenishment before final sales fall by the same amount.
Seasonality, Revisions and Release Lags
Comparisons need matching periods and seasonal treatment. In December 2023, Cass shipments fell 1.6% month to month but rose 2.1% on the report’s seasonally adjusted comparison. Activity fell less than the usual seasonal pattern; the two columns answer different questions. [13]
BTS’s comparison describes the core Cass series as not seasonally adjusted, whereas individual Cass reports also provide labeled adjusted estimates. A workday adjustment alone is not the same as removing all seasonal effects. Check the specific column, rather than assigning one treatment to everything carrying the same provider name. [4]
Freight TSI inputs are seasonally adjusted, and BTS can forecast a delayed input. It revises the index when inputs, historical methods or estimated seasonal patterns change. Federal Reserve industrial production is preliminary and also undergoes source-data and benchmark revisions. Its October 16, 2026 supplement, for example, showed later-month blanks for the detailed paperboard-container series even though other rows had newer observations. A current release date does not mean every detailed row describes the same month. [15] [8]
For PPIs, industry-classified indexes are published unadjusted; selected commodity indexes also have adjusted versions. Initial price observations can be revised, so retain the release vintage. For AF&PA, its annual time-series product explicitly distinguishes adjusted and unadjusted monthly histories; a short quarterly news release is not a substitute for all of that documentation. Do not assume a year-over-year comparison removes every calendar, composition or base effect. [5] [6] [10]
Historical Divergences and False Recession Signals
A narrow contraction can be correctly measured while an economy-wide recession inference is wrong. Two historical comparisons show why the distinction matters.
2019: freight weakness during a continuing expansion
Cass reported December 2019 shipments 7.9% below December 2018. Yet the NBER subsequently dated the monthly economy-wide expansion as continuing through February 2020. Treating the freight decline as proof that a monthly U.S. recession was already underway in 2019 would have been incorrect. The weakness in covered freight was real; the national conclusion required broader evidence. [11] [12]
This does not establish that the series had no advance-warning value. A later pandemic recession occurred, and NBER’s quarterly peak was 2019 Q4 even though its monthly peak was February 2020. Testing a forecast requires specifying its rule and horizon beforehand; assigning a convenient lead time afterward is not evidence of accuracy.
2023: declining freight alongside annual GDP growth
Cass reported that its shipment index fell 5.5% for 2023 and its expenditure index fell 19%. In its January 25, 2024 release, BEA estimated that annual-average real U.S. GDP increased 2.5% in 2023. That dated GDP estimate includes much broader production than the freight sample. Later GDP revisions do not turn these different concepts into interchangeable measures. [13] [14]
Annual GDP growth alone does not prove that every month or quarter expanded, and this comparison is not a complete recession-dating test. It does establish that a substantial decline in covered freight can coexist with growth in annual economy-wide real output. The NBER’s framework examines the breadth of economic contraction, rather than defining a recession from one industrial series. [12]
How to Use Freight and Packaging Data
Match the series to the decision first. For a carrier, a relevant volume measure and a matching price measure can help frame revenue pressure, but fuel, labor and other costs still determine margins. For a packaging producer, output, trade exposure and plant capacity can matter more than an undifferentiated national freight count.
For shipping-company analysis, time charter equivalent addresses a different question: how voyage revenue after specified voyage expenses translates into a rate per vessel day. The day denominator and issuer reconciliation matter; the rate is not company profit.
For a warehouse investment, national freight data provides context; local tenant demand, routes, new construction and vacancies remain essential. Our real estate market analysis guide explains that geographic and property-type boundary.
For a macro conclusion, use business cycle indicators that observe other parts of production, employment, income and spending. Several freight measures sharing clients, modal data or packaging sources are not necessarily independent confirmations. Record the series’ name, covered population, unit, last observation, seasonal treatment and release vintage alongside the conclusion it supports.
A defensible statement might be: “Covered shipment activity weakened, while spending fell faster; composition and service-price data are needed to explain the difference.” Moving from that observation to a national forecast requires the broader framework discussed in probability of recession, not a universal freight threshold.
Common Mistakes
- Calling spending growth volume growth. Separate movements, service prices and composition before interpreting a larger freight bill.
- Calling expenditures divided by shipments a pure price index. The average can change when the service mix changes even if each specified price stays fixed.
- Comparing different index levels. A TSI level of 100 and a Cass level of 1.00 reflect different reference bases and constructions, not equal quantities.
- Equating mill tonnage with box counts or domestic sales. Check the production stage, physical unit, inventory and trade boundary.
- Mixing raw and adjusted monthly changes. Use the actual series/column labels and match periods before drawing a trend.
- Backtesting with revised information as if it were available then. Save the original releases and define the prediction horizon before assessing a signal.
Limitations of Freight and Packaging Indicators
Freight and packaging data can identify genuine industrial stress without establishing a broad recession. They also cannot determine a stock’s fair value, a carrier’s profitability or a warehouse’s local demand on their own.
Public aggregates usually cannot isolate every change in customer coverage, shipment size, route, contract, inventory behavior or transport substitution. Seasonal models and revisions add uncertainty. A newer observation is useful only if its construction fits the question, and two historical divergences do not supply a calibrated forecast probability or a general false-positive rate.
Use these measures to sharpen questions about physical activity, pricing and industrial conditions. Require separate evidence for the broader economic conclusion and for the financial decision that follows.
Frequently Asked Questions
Sources
- Cass Information Systems: Cass Freight Index methodology. Client coverage, shipment counts, expenditures, processing timing and reference base.
- BTS: Methodology for the Transportation Services Index, November 3, 2025. Physical inputs, weights, seasonal adjustment and chaining.
- BTS: Transportation Services Index FAQs. Covered modes, exclusions and limitations.
- BTS: What the Transportation Services Index, Dow Transportation Index, and Cass Freight Index Tell Us. Construction and timing differences.
- BLS: PPI Guide for Price Adjustment. Sampling, scope, index selection and preliminary/final observations.
- BLS: PPI Concepts and PPI FAQs. Specified prices, interpretation and seasonal treatment.
- Federal Reserve: G.17 Source and Description Table 1.19, Paper, November 24, 2025. Paperboard mill and container source units.
- Federal Reserve: Industrial Production Explanatory Notes and October 16, 2026 Table 1B supplement. Output estimates, revisions, seasonality and detailed-series availability.
- AF&PA: Q1 2024 Containerboard Quarterly Report public release, April 26, 2024. Production, exports and domestic new supply.
- AF&PA: 2024 Paper Production and Capacity Trends release and Containerboard Annual Time Series description. Public survey coverage and adjusted/unadjusted product description; paid tables not used.
- Cass: December 2019 Transportation Index Report. Dated shipment decline.
- NBER: Determination of the February 2020 Peak in Economic Activity, June 8, 2020. Monthly/quarterly dates and economy-wide definition.
- Cass: December 2023 Transportation Index Report. Annual changes, seasonal comparison and inferred-rate limitations.
- BEA: Gross Domestic Product, Fourth Quarter and Year 2023, January 25, 2024 vintage. Annual real GDP growth and components.
- BTS: Revision Policy for the Transportation Services Index, November 22, 2024. Source, methodology and seasonal revisions.
- BLS via FRED: General Freight Trucking, Long-Distance Truckload PPI, PCU484121484121. Official series identity, base, frequency and seasonal label; no current level used.
Sources accessed November 15, 2026 UTC. Historical figures retain their stated observation periods and release vintages. Both shipper examples are hypothetical author calculations.
Disclaimer
This article is for educational and informational purposes only and does not constitute investment or business advice. Freight and packaging measures have limited coverage and can be revised. Historical relationships and hypothetical calculations do not predict economic turning points, company results or investment returns. Consult a qualified professional about your circumstances.
