What are your input costs doing, and where will they be in 12 months?

For auto and ev manufacturing (NAICS 3361), a weighted basket of World Bank Pink Sheet commodity prices sits at an index of 241.3 in 2026M01, with 2016M01 = 100. That is +141.3% over the last 10 years, with Iron ore, cfr spot the largest single contributor to the net move. The Holt linear-trend forecast carries the basket to 236.7 in 12 months (-1.9%), with an 80% interval of [170.0, 329.5]. Other sectors: construction · electronics manufacturing · food processing · apparel manufacturing ·

Basket composition and weight provenance

The weights below are taken from the published references listed in each sector block, not imputed from the BACI trade flows or fitted to any model. Where a sector uses an input that the World Bank Pink Sheet does not quote directly (for example, flat steel, engineering plastics, or polyester fibre), the row is flagged as an upstream proxy. This is a substantive limitation of using only open reference prices: a refined-material index would track the downstream input more tightly, at the cost of licensing a paid data feed (CRU for steel, ICIS for plastics, PCI Wood Mackenzie for fibre).

Weight source: ICCT (Lutsey 2017) Table 3 material cost shares; McKinsey (2019) BEV bill-of-materials; BloombergNEF (2023) EV Outlook material intensity. Proxy convention documented in method note.

How the basket has moved over 10 years

Figure 1 plots the weighted input-cost index for auto and ev manufacturing. The index uses Tornqvist-style geometric aggregation across the basket, I(t) = exp(sum_i w_i ln(p_i(t) / p_i(t0))) times 100, which is the aggregation used by the IMF Primary Commodity Price System and is closely related to the modified Laspeyres used by BLS PPI (BLS Handbook of Methods, Chapter 14). Geometric aggregation dampens the arithmetic sensitivity to a single commodity spike while preserving substitution-neutral elasticity at the basket level.

Commodity-by-commodity contribution

Figure 2 decomposes the basket into one line per commodity, where each point is the commodity’s weight times its price ratio to 2016M01, times 100. This is the arithmetic contribution reporting convention documented in BLS Handbook of Methods, Chapter 14 (“Contributions to change”). Reading across the lines shows which single inputs the basket’s movement is most sensitive to. Summed arithmetically, these lines recover the Laspeyres level; the geometric basket in Figure 1 is the log-weighted alternative.

12-month forecast with 80% confidence band

Figure 3 extends each basket commodity with a Holt linear-trend exponential smoother on log prices, then re-aggregates to the basket with the same weights. The method is the one documented in Hyndman and Athanasopoulos (2021), Forecasting: Principles and Practice, 3rd ed., Chapter 8.3 (Holt’s linear method) and 8.7 (prediction intervals), available at otexts.com/fpp3. The original references are Holt (1957), “Forecasting Seasonals and Trends by Exponentially Weighted Moving Averages”, ONR Memo 52, and Winters (1960), “Forecasting Sales by Exponentially Weighted Moving Averages”, Management Science 6(3): 324–342. Smoothing parameters: alpha = 0.3, beta = 0.1, the conservative default in HA Table 8.10 for commodity-style series. The basket 80% interval is computed in log space as z_{0.80} times sqrt(h * sum w_i^2 sigma_i^2), which assumes independent commodity innovations; co-movement (copper-aluminum, oil-gas) widens the true interval relative to the one shown.

Shipping-cost overlay

Input costs landed at the factory gate are the sum of the commodity basket and the freight cost to move it. Figure 4 overlays two published freight indices onto the auto and ev manufacturing basket: the BLS Producer Price Index for Deep Sea Freight Transportation (series PCU4831114831115, monthly since 1988) and the Cass Freight Index (series FRGSHPUSM649NCIS, monthly since 1990), both sourced from FRED. Baltic Exchange Dry Index and Freightos Baltic Index (FBX) are the purer daily spot-freight references but are not yet ingested into the workbench (TODO: add Baltic Exchange and Freightos feeds once licensed); the two FRED series above are the best-in-workbench proxies for now, and both are primary-source published indices, not modelled.

Cross-sector comparison, latest month

Figure 5 plots the latest-month basket index level for each of the five sectors against the same 10-year window = 100 base. The ranking reflects differences in material-cost structure: a copper- and gold-heavy electronics basket has different price memory than a cotton-and-oil apparel basket. Production-network economics (Carvalho and Salehi 2019, “Production Networks: A Primer”, Annual Review of Economics 11: 635-663; Baqaee and Farhi 2024, “Networks, Barriers, and Trade”, Econometrica92(2): 505-541) imply that sector-level price shocks propagate along input-output linkages with multipliers above one, so the gap between the leader and the laggard sector understates the within-firm pass-through seen by a final producer.

Pass-through coefficient by industry: energy-cost elasticity of input baskets

Carvalho and Salehi (2019, “Production Networks: A Primer”, Annual Review of Economics 11: 635-663) show that upstream input-price shocks propagate to downstream sectors with multipliers determined by the Leontief inverse of the input-output matrix. The full Leontief decomposition requires a BEA / OECD-ICIO style table, which is not wired into this workbench; Figure 6 instead computes a bounded empirical proxy: the pass-through elasticity beta_k of each sector-k basket to the Brent crude oil price, estimated by OLS regression of dlog(basket_k,t) on dlog(Brent_t) on monthly data over the 10-year window in Figure 1. Brent is the pivot commodity: Acharya, Berner, Engle, Jung, Stroebel, Zeng & Zhao (2023, “Climate Stress Testing”, Annual Review of Financial Economics 15: 291-326) document that energy-price shocks account for the plurality of input-cost volatility across G-10 manufacturing sectors, so the Brent elasticity is the canonical first-moment pass-through channel. Amiti, Itskhoki & Konings (2019, “International Shocks, Variable Markups, and Domestic Prices”, Review of Economic Studies 86(6): 2356-2402) frame the interpretation: a pass-through beta above 1 indicates cost amplification through intermediates; below 1 indicates substitution or hedging buffers.

Pass-through speed: months until 50% of an energy-price shock hits the basket

The Figure 6 beta gives the total cumulative energy-price pass-through; this figure gives the speed at which it arrives. For each sector-k basket, an ordinary least-squares distributed-lag regression of dlog(basket_k,t) on dlog(Brent_{t-L}) for L=0,1,...,12 recovers the lag weights beta_L. Cumulating beta_L up to lag L and dividing by the sum gives the cumulative pass-through fraction; the smallest L at which the fraction exceeds 0.5 is the pass-through half-life. Gopinath, Itskhoki & Rigobon (2010, “Currency Choice and Exchange Rate Pass-Through”, American Economic Review 100(1): 304-336) document that half-life estimates for imported-input pass-through concentrate in the 3-9 month range across advanced-economy industries; the workbench estimates below place each basket inside that range.

Terms-of-trade for resource-rich countries vs manufacturers

Figure 8 compares two bundles of Pink Sheet commodities on a common axis, using price_real (deflated by the World Bank Manufactures Unit Value index per Pink Sheet Methodology Note, 2023) so each line is already an implicit terms-of-trade series against manufactures. The resource-extraction bundleis an equally-weighted geometric mean of energy and industrial metals (Brent, US natural gas, copper, aluminum, iron ore); the manufactures-input bundle is an equally-weighted geometric mean of agricultural and lighter-industry feedstocks (wheat HRW, cotton A Index, rubber RSS3, sawnwood Malaysian). A rising resource bundle relative to the manufactures bundle is a positive terms-of-trade shock for resource-rich exporters (the Prebisch-Singer reversal documented by Cuddington & Jerrett, 2008, and Erten & Ocampo, 2013); a rising manufactures-input bundle squeezes downstream manufacturing margins (Acharya et al., 2023, Annual Review of Financial Economics 15: 291-326).

Which inputs are noisiest: annualised volatility of each basket commodity

Carryable across all the figures above is one practical question for a procurement team: of the 6 commodities in this sector basket, which ones bring the most month-to-month price noise into the cost line, and therefore deserve the first-priority forward cover or supplier re-negotiation? Figure 9 answers it directly: for each basket commodity, we compute the standard deviation of monthly log returns over the 10-year window and annualise by sqrt(12), the convention in Mandelbrot & Hudson (2004) and the Hull (2018, Options, Futures, and Other Derivatives, 10th ed.) treatment of commodity volatility. Higher bars are noisier inputs. The ranking is independent of the weight: a small-weight but high-volatility commodity (say, natural gas in food processing) can still drive the bulk of unhedged P&L variance through Bohi-Toman (1996) and the Acharya et al. (2023) energy-shock channel. Multiplying volatility by basket weight gives the variance contribution to the basket and is the canonical hedging-priority metric (the contribution-to-risk decomposition in Litterman 1996, Goldman Sachs Risk Management Series).

How CFOs and procurement teams use this

• Scenario planning. The Figure 3 upper-bound path is the hedging-trigger scenario; the lower bound is the procurement-flexibility scenario. Budget against the central path but stress-test against both tails.
• Contract renegotiation. When Figure 2 shows one commodity dominating basket movement, renegotiate the underlying contract on that commodity first. Weighted exposure dictates hedging priority.
• Incoterms choice. When Figure 4’s freight lines decouple above the basket, switch from CIF-style landed-cost contracts to FOB + spot-freight arbitrage. When they decouple below, lock FOB.

Method and data notes

• Weights are from published sources, listed in the basket table for each sector. We do not fit weights from the price data, which would introduce look-ahead bias.
• Upstream proxies are flagged. A refined-material index would track the downstream input more tightly but requires paid feeds (CRU for steel, ICIS for plastics, PCI Wood Mackenzie for fibre). The proxy convention is the honest compromise for an open workbench.
• Pink Sheet is nominal USD. The index is in nominal terms; to deflate, divide by the US CPI or BLS PPI for final goods in the sector.
• Holt linear over Holt-Winters. Monthly commodity series in the Pink Sheet show weak and inconsistent seasonality; imposing a 12-month seasonal component tends to over-fit to energy winter spikes. See HA Ch. 8.5.
• Independent-innovation CI. The 80% band assumes independent commodity shocks. Copper-aluminum and oil-gas co-move positively, so the true CI is wider than stated. Treat the band as a lower bound on uncertainty.

Policy read: how 2024-2026 climate and trade policy propagates into input costs

The five sector baskets are all exposed to the same four policy shadows. COP29 (Baku, November 2024) raised the implied carbon cost of imported energy-intensive inputs via Article 6 operationalisation; this transmits first into steel and aluminum for the auto and construction baskets. The EU Fit for 55 package (COM(2021) 550) pairs CBAM with ETS2 for road and building fuels from 2027 and layers the Methane Regulation (Reg 2024/1787) on gas imports, which raises the natural-gas line in the auto and food-processing baskets. The US Inflation Reduction Act (Public Law 117-169, 2022) subsidises domestic battery-material supply (lithium, nickel, cobalt) under section 45X, which damps the copper and nickel lines of the electronics basket with a lag. The WTO Agreement on Agriculture Article 12 and the MC12 Ministerial Decision on Food Export Prohibitions bound the cereal-exporter policy space that drives the food-processing basket’s tail risk; Carvalho-Salehi (2019) and Baqaee-Farhi (2024) formalise how these origin-country shocks amplify through the production-network into final goods prices.

References

• Acharya, V. V., Berner, R., Engle, R., Jung, H., Stroebel, J., Zeng, X., & Zhao, Y. (2023). “Climate Stress Testing.”Annual Review of Financial Economics 15: 291-326.
• Baqaee, D. R., & Farhi, E. (2024). “Networks, Barriers, and Trade.” Econometrica 92(2): 505-541.
• Carvalho, V. M., & Salehi, A. (2019). “Production Networks: A Primer.” Annual Review of Economics 11: 635-663.
• Bureau of Labor Statistics (2023). Handbook of Methods, Chapter 14: Producer Price Indexes. https://www.bls.gov/opub/hom/pdf/homch14.pdf
• Hyndman, R. J., & Athanasopoulos, G. (2021). Forecasting: Principles and Practice, 3rd ed. OTexts, Melbourne. https://otexts.com/fpp3/
• Holt, C. C. (1957). “Forecasting Seasonals and Trends by Exponentially Weighted Moving Averages”. Office of Naval Research Memorandum No. 52.
• Winters, P. R. (1960). “Forecasting Sales by Exponentially Weighted Moving Averages”. Management Science 6(3): 324–342.
• Lutsey, N. (2017). “Modernizing vehicle regulations for electrification”. International Council on Clean Transportation (ICCT) White Paper.
• McKinsey & Company (2019). “Unlocking the full potential of vehicle electrification”.
• BloombergNEF (2023). Electric Vehicle Outlook 2023.
• USGS (2024). Mineral Commodity Summaries 2024. U.S. Geological Survey.
• USDA Economic Research Service (2023). Commodity Costs and Returns; Food Dollar Series.
• OECD (2021). “The Apparel Industry and Its Inputs”. OECD Trade Policy Papers.
• World Bank (monthly). Pink Sheet: Commodity Markets Monthly Data. https://www.worldbank.org/en/research/commodity-markets
• International Monetary Fund. Primary Commodity Price System (PCPS) methodology.
• Amiti, M., Itskhoki, O., & Konings, J. (2019). “International Shocks, Variable Markups, and Domestic Prices”. Review of Economic Studies 86(6): 2356–2402. (Cost pass-through context.)