Natural Gas Outlook: Key Factors Driving Future Prices

1. The Global Supply-Demand Tightrope: LNG Flows and Storage Dynamics
The interplay between liquefied natural gas (LNG) supply and global storage inventories remains the most immediate driver of price volatility. The post-2022 era reshaped the market, making European and Asian storage levels far more critical than in previous decades. For the near-term outlook, watch the injection season (April–October in the Northern Hemisphere). If storage fill rates lag due to a cold spring or early-season heatwaves boosting cooling demand, spot prices for TTF (Title Transfer Facility) and JKM (Japan Korea Marker) will spike, dragging Henry Hub higher via LNG arbitrage channels. Conversely, a mild winter combined with robust US export capacity could lead to storage overfills, pressuring prices downward. Key data points: the EIA’s Weekly Natural Gas Storage Report, GIE’s (Gas Infrastructure Europe) storage dashboard, and LNG cargo tracking data from Kpler or Vortexa. Structural supply constraints—such as delays in Freeport LNG’s expansion and the gradual phase-out of Russian pipeline flows—mean that any marginal demand shock will have an outsized impact on front-month futures.

2. The Weather Wild Card: Degree Days and HVAC Electrification
Weather variability accounts for roughly 50–60% of short-term residential and commercial demand in the US. The transition from winter heating to summer power generation creates a dual-threat sensitivity. In the US, the National Oceanic and Atmospheric Administration (NOAA)’s three-to-six-month outlooks are scrutinized for Heating Degree Day (HDD) and Cooling Degree Day (CDD) anomalies. An anomalously hot summer (CDDs above the 10-year average) directly boosts gas-fired power plant generation as air conditioning loads surge—especially as coal retirements accelerate. A colder-than-average winter (HDDs spiking in the Midwest and Northeast) can drain storage rapidly, triggering Henry Hub winter strip prices above $5/MMBtu. A significant emerging factor: the electrification of heat pumps. As these systems become more efficient, they reduce winter gas demand per home but increase overall grid reliance, potentially creating new, less predictable demand peaks during polar vortex events.

3. US Production Growth vs. Operator Discipline: The EIA 914 Report
The near-record dry natural gas production in the Permian Basin and Haynesville Shale has reshaped US fundamentals, but operator capital discipline is capping growth. The EIA’s monthly 914 production report and the Baker Hughes rig count reveal a flattening trajectory: even at $2.50–$3.00/MMBtu Henry Hub, many drillers prefer returning cash to shareholders via dividends and buybacks rather than adding rigs. The Permian now produces roughly 40% of total US associated gas, making production inextricably linked to oil prices. A $10/barrel drop in WTI can reduce gas output by 0.5–1.0 Bcf/d as operators shut-in low-IP wells. The Northeast (Appalachia) faces pipeline takeaway constraints, capping basis differentials at Dominion South and TCO. The key leading indicator: producer hedging activity. If hedges expire without being replaced at current prices, production growth could stall further. The structural decline in dry gas-directed drilling (Haynesville rigs down 40% from 2023 peaks) suggests a production plateau near 102–104 Bcf/d through Q3 2025.

4. The Power Sector Transition: Coal Retirements and Renewables Intermittency
The US power sector burns approximately 30–35 Bcf/d of natural gas, making it the single largest demand segment. Coal-to-gas switching economics remain a dominant price floor. When Henry Hub falls below $2.00/MMBtu and coal prices remain above $2.50/MMBtu, utilities dispatch gas over coal almost universally. This creates a natural price cap on downward movement during shoulder seasons. However, the rapid deployment of solar and wind generation (now exceeding 25% of total US capacity in ERCOT and CAISO) introduces a new volatility driver: the “duck curve.” During solar-rich spring afternoons, gas generation can drop below 15 Bcf/d, only to spike above 30 Bcf/d during evening ramps. This intraday volatility is not fully captured in monthly forward curves, but it affects storage injection schedules and pipeline operational flow orders (OFOs). For the 2025–2026 outlook, expect a 2–4% annual increase in total gas demand from power, driven by data centers and AI compute loads (see Section 6).

5. The Data Center Demand Surge: AI, Cloud, and the Decarbonization Tension
The single most transformative long-term demand driver is the projected 20–30 GW increase in US data center power demand by 2030. This is equivalent to adding another New York City to the grid every two years. Natural gas is the most scalable bridge fuel to meet this demand, as new nuclear and geothermal facilities face decade-long permitting timelines. Major tech players (Google, Microsoft, Amazon) have signed long-term virtual power purchase agreements (VPPAs), but they are also negotiating direct-feed gas interconnection agreements with utilities for gigawatt-scale campuses in Virginia, Ohio, and Arizona. The impact on natural gas prices is twofold: (1) it increases total structural demand by 3–5 Bcf/d by 2028, tightening the supply-demand balance; (2) it creates regional basis blowouts at hubs like Dominion South, Henry Hub, and Waha (West Texas), as new load centers outpace pipeline capacity. For the 2025–2026 window, the North American Electric Reliability Corporation (NERC) warns of capacity shortfalls in PJM and MISO, which will likely drive winter peak gas prices significantly higher than summer strips.

6. Liquefied Natural Gas (LNG) Export Capacity and Geopolitical Leverage
US LNG exports have transformed from a marginal outlet to a structural demand driver, consuming roughly 12–14 Bcf/d of domestic supply in 2024 and projected to reach 16–18 Bcf/d by 2027 with the completion of Plaquemines LNG, Corpus Christi Stage 3, and Golden Pass (pending final investment decisions). Each new LNG train creates a hard price floor for Henry Hub, as export buyers are price-insensitive relative to domestic consumers (LNG shipped to Asia can sustain $12/MMBtu TTF equivalents while Henry Hub remains at $3). Key geopolitical factors influencing this flow: (1) the US Department of Energy’s pause on new non-FTA export permits—any resumption or amendment will affect 2028+ forward curves; (2) the energy crisis in Europe and Asia’s scramble for long-term contracts post-Ukraine; (3) Panama Canal drought restrictions, which increase shipping costs and redirect cargoes to the Suez or around Cape Horn, altering regional pricing spreads. The JKM-TTF-Henry Hub basis is the single most important inter-market signal for price direction. When TTF futures exceed $15/MMBtu, US LNG flows become mechanically saturated, creating a “virtual pipeline” floor for US prices.

7. Structural Pipeline Constraints and Regional Basis Divergence
The US gas transportation network is aging, and bottlenecks are amplifying price disparities. Waha Hub (Permian Basin) frequently trades at a $1.50–$2.00/MMBtu discount to Henry Hub due to insufficient pipeline takeaway capacity to the Gulf Coast. Any maintenance at the Gulf Coast Express or Permian Highway Pipeline can send Waha negative. Conversely, New England (Algonquin Citygate) suffers from severe winter basis blowouts exceeding $10/MMBtu due to pipeline constraints and limited LNG import terminals (Everett). The Northeast’s lack of new pipeline buildout (due to regulatory hurdles) means that winter 2025–2026 basis risk is exceptionally high. For the outlook, focus on FERC pipeline dockets, shipper contract renewals, and expansion projects like the MVP Southgate and the Permian-to-Gulf expansions (Matterhorn Express). A sustained basis blowout in a consuming region signals a broader national supply shortage, pushing Henry Hub futures higher as traders hedge against delivery risk.

8. Environmental Regulations and the Methane Fee
The US Environmental Protection Agency (EPA)’s Waste Emissions Charge (methane fee), established under the Inflation Reduction Act, imposes a $900–$1,500/ton fee on emissions exceeding statutory thresholds, phased in from 2025. For large producers, this adds an estimated $0.05–$0.10/MMBtu to production costs—material enough to tip marginal wells toward shut-ins. Furthermore, the EPA’s updated methane rule (40 CFR Part 60) requires continuous monitoring and leak detection, adding operational costs and reducing flow assurance. In states like Colorado and New Mexico, additional state-level rules (Colorado O&G Regulation 900, New Mexico’s Venting and Flaming rules) already curtail production growth. These regulatory headwinds effectively tighten the forward supply curve, especially for high-emission legacy production in the Permian and Bakken. Traders should monitor EPA enforcement actions and state compliance reports for signs of accelerated well abandonment.

9. The Structural Decline of European Gas Production and Russian Flows
European indigenous output—particularly from the Dutch Groningen field (now permanently closed) and the UK North Sea—is in terminal decline, falling by 5–7 Bcf/d per year. Simultaneously, Russian pipeline flows to Europe via Ukraine have contracted to near zero after the expiration of the transit agreement in December 2024. This leaves Europe almost entirely reliant on LNG for marginal supply. Any disruption to Norwegian infrastructure (maintenance at Nyhamna or Karstø) or an extended cold spell in Europe immediately tightens the global LNG market, sending TTF above $20/MMBtu and pulling US LNG cargoes from Asia. The result: a direct transmission of European weather and geopolitical risk into Henry Hub. The US is no longer a closed economy; it is the swing supplier to a structurally deficit Atlantic Basin. For the 2025–2027 outlook, assume a baseline TTF of $12–$18/MMBtu, which mechanically supports a Henry Hub floor of $2.80–$3.50/MMBtu.

10. The Economic Recession and Industrial Demand Risk
Industrial demand (fertilizer, chemicals, refining) consumes roughly 18 Bcf/d in the US. A sharp US recession (GDP contraction >1%) would reduce industrial output by 2–4%, translating to a 0.5–1.0 Bcf/d demand loss. However, this is partially offset by the fact that US industrial facilities (particularly ammonia and methanol plants) have already rationalized production at high gas prices. The more significant risk is a global recession, which would reduce LNG demand in Europe and Asia, collapsing TTF and allowing cargoes to flood back into the Atlantic Basin. This would create a tail-risk scenario where Henry Hub could test $1.50/MMBtu intra-month. Conversely, a “soft landing” (inflation moderating, employment stable) supports demand growth, as new ethylene crackers and blue hydrogen projects continue construction. The Federal Reserve’s rate decisions and Purchasing Managers’ Index (PMI) readings are therefore critical leading indicators for gas demand beyond the power sector.

11. Storage Fill Trajectories and the December De-Annualization Trade
The most actionable near-term price catalyst is the pace of U.S. working gas storage inventories relative to the five-year average. As of late Q1 2025, storage sat near 1.8 Tcf (above the five-year average), creating a bearish near-term bias. However, the “Lag Time Effect” of production declines means that if storage injections from April to June fall 10 Bcf/week below normal, the market could flip to a deficit by August. The critical level: if inventories end the injection season below 3.6 Tcf (five-year average is 3.8 Tcf), the winter price strips will re-price sharply higher. The “De-Annualization Trade”—selling monthly strips against Dec/Jan futures—becomes profitable when storage is tight. Track the week-over-week EIA storage report, particularly the South Central region, where salt dome storage accounts for most of the flexibility. A storage deficit in this region is a direct bullish signal for the full gas complex.

12. The 2025–2026 Hurricane Season and Gulf Production
The Gulf of Mexico accounts for roughly 2–3 Bcf/d of dry gas production, plus critical processing and LNG infrastructure. The NOAA’s hurricane outlook (issued in May) forecasts an above-average season (12–17 named storms). A direct hit on the Louisiana coast near Sabine Pass or Cameron LNG can shut in 1-2 Bcf/d of production and disrupt LNG loading for 7–14 days. The forward curve for Q3 2025 already bakes in a modest hurricane risk premium of $0.15–$0.25/MMBtu. However, a single category 3+ storm making landfall near Port Arthur or Lake Charles can send the month-ahead contract vertical, resembling August 2020 (Hurricane Laura). For active traders, monitoring the Atlantic Basin tropical wave activity and the GoM-specific shut-in data (BSEE) is non-negotiable. The 2025 season may be particularly volatile due to elevated sea surface temperatures and the transition from El Niño to La Niña.

13. Technological Disruption: Green Hydrogen and Carbon Capture Feedback Loops
While green hydrogen remains nascent, the tax credits under 45V (Clean Hydrogen Production Tax Credit) are incentivizing natural gas-based blue hydrogen projects with carbon capture and storage (CCS). Each large blue hydrogen facility (e.g., CF Industries’ Donaldsonville expansion, ExxonMobil’s Baytown blue hydrogen plant) consumes 0.2–0.5 Bcf/d of natural gas as feedstock and fuel. These projects create a direct, long-term demand sink, even as policymakers push for decarbonization. Conversely, successful CCS development could allow low-carbon gas to compete with renewables for power generation, sustaining gas demand into the 2030s. The regulatory status of 45V’s “additionality” requirements (if and when finalized) will determine whether these projects proceed. For now, blue hydrogen remains a bullish undercurrent for long-dated gas futures.

14. Futures Curve Structure, Basis Swaps, and Hedging Activity
The NYMEX Henry Hub forward curve provides the most visible consensus price outlook. A frequently observed structure in 2025 is a contango from spring to winter months, reflecting the cost of carry, followed by a backwardated winter peak. The forward spread between the April-September strip and the November-December strip (the “winter risk premium”) is a direct measure of market anxiety about storage adequacy. In addition, the proliferation of basis swaps (e.g., Waha/Henry Hub, TTF/Henry Hub, Dominion South/Henry Hub) enables producers and utilities to hedge regional risks. A sustained widening of the TTF/Henry Hub spread above $5/MMBtu indicates tightening Atlantic Basin conditions and has historically preceded a 5-8% move up in Henry Hub. For granular outlook, monitor the Commitment of Traders (COT) report—if managed money (speculators) are net short and commercial hedgers are net long, it often signals a top in price.

15. Policy Shocks: The DOE LNG Export Pause and Carbon Border Taxes
The Biden administration’s pause on new non-FTA LNG export approvals (announced January 2024) created immediate uncertainty for projects targeting 2027–2029 in-service dates (e.g., Venture Global’s CP2, NextDecade’s Rio Grande expansions). A reversal of this pause—either by the current administration (if challenged in court) or a new administration after the 2024 election—would be a major bullish catalyst for long-dated Henry Hub, as it signals commitment to US LNG as a strategic resource. Conversely, a permanent ban or heavy restrictions would cap long-term demand growth, flattening the far end of the curve. Additionally, the implementation of Europe’s Carbon Border Adjustment Mechanism (CBAM) for imported LNG, starting in 2026, would impose a cost on high-emissions US LNG cargoes. US producers with high methane emissions (proxied by the IOGP’s methane intensity data) would face a $0.20–$0.50/MMBtu cost disadvantage, shifting buying patterns toward lower-intensity supplies (e.g., Qatar, Australia). This policy divergence will create tiered pricing for global LNG.

16. The Role of Financial Speculation and Algorithmic Trading
Open interest in NYMEX Henry Hub futures has grown significantly, with algorithmic traders now accounting for roughly 60–70% of daily volume via CME Globex. This has compressed intra-day Sharpe ratios but amplified intra-week momentum. A surprise storage release (e.g., a 35 Bcf injection vs. 55 Bcf consensus) can trigger a sudden 20-cent move in minutes due to cascading stop-loss orders. For the longer-term outlook, the net managed money position is a contrarian indicator—when it reaches extreme levels (e.g., net short >60k contracts), the market tends to reverse violently. Traders should overlay the Commodity Futures Trading Commission (CFTC) weekly COT data with the EIA’s Short-Term Energy Outlook (STEO) to gauge whether the speculative consensus is aligned with physical fundamentals. The 2025 outlook will be particularly prone to flash crashes and rallies as the market digests conflicting signals from production and LNG demand.

17. Natural Gas Storage Optimization: Salt Domes vs. Depleted Reservoirs
The physical characteristics of US natural gas storage facilities are an overlooked price driver. Roughly 20% of capacity is in high-deliverability salt dome storage (primarily in the Louisiana Gulf Coast), with withdrawal rates of up to 1 Bcf/d per facility. During extreme cold events, these sites can provide emergency supply; when they run low, the market loses its last line of defense. Conversely, depleted reservoir storage (in the Midwest and Northeast) has lower withdrawal rates but larger total capacity. A season when salt dome inventories drop below 40% of capacity by January is a clear warning sign of price spikes, as only these sites can meet peak-day demand. For the 2025–2026 winter, if salt dome storage exits injection season below 200 Bcf (total capacity ~400 Bcf), expect winter prices to reach $6–8/MMBtu on New York Mercantile Exchange close to cold spells.

18. The Decarbonization Paradox: Gas as a “Transition Bridge” to Renewables
Policymakers in the EU, US, and Asia increasingly recognize that natural gas is necessary to stabilize grids while battery storage scales. The International Energy Agency (IEA) projects that global gas demand will increase by 3–5% through 2030 in the “Stated Policies Scenario,” driven by emerging markets and data centers. This narrative supports long-term investment in gas infrastructure. However, the “Net Zero by 2050” scenario suggests a plateau and decline after 2035, creating a pricing tension: near-term demand is structurally underpinned, but the 2035+ overhang weighs on long-dated contracts. This “duration risk” means that the 2026–2028 futures strip (DEC28) often trades at a discount to the first-year contracts, reflecting investor pessimism about the long-run. This creates opportunities for calendar spreads: buy near-term, sell long-term, capturing the “green premium” erosion.

19. Demographic and Behavioral Shifts in Demand
Household natural gas consumption per customer has declined by roughly 15% over the past decade due to energy-efficient appliances, better building insulation, and LED lighting. This structural decline in per-capita use of residential gas is partially offset by population growth and new housing starts. However, a new wildcard is the “work-from-home” effect: as employees continue to work from home more days per week than pre-pandemic, residential winter demand remains elevated compared to commercial office building demand. This distributional shift favors winter gas demand over shoulder months. The EIA’s monthly Residential Energy Consumption Survey (RECS) provides this granular data.

20. Cross-Commodity Correlations: Oil, Coal, and Carbon Prices
Natural gas is increasingly correlated with European carbon allowance prices (EUA) and global coal prices, particularly for and beyond the 12-month forward time horizon. When coal prices exceed $120/ton in Europe, gas-to-coal switching becomes uneconomical, boosting gas demand. Conversely, a collapse in carbon prices (below €60/ton EUA) reduces the cost advantage of gas over coal in European power markets. In the US, the correlation between Henry Hub and WTI oil is currently weak (roughly 0.2) due to the decoupling of oil production from direct gas demand, but a sudden oil price spike pushes up domestic heating oil prices, making gas more attractive. Traders must watch the gas-to-coal spread (NYMEX NG vs. API2 coal) and the TTF-to-carbon spread for early signals of fuel-switching behavior.

21. The Role of the US Strategic Petroleum Reserve and Its Spillover
The US Strategic Petroleum Reserve (SPR) is not directly linked to gas, but its drawdowns and reinjections affect crude oil markets, which influence gas via production economics and investor sentiment. When the Department of Energy refills the SPR (via fixed-price contracts), it supports WTI and thus associated gas production. Conversely, a large SPR drawdown to combat high oil prices reduces associated gas supply. While the impact is small (0.1–0.3 Bcf/d), it is additive. The DOE’s refill schedule (typically 1–3 million barrels per month) is a minor bullish factor for dry gas production in the Permian.

22. Price Elasticity of Demand: The Self-Correcting Mechanism
One of the underappreciated features of the natural gas market is the negative price elasticity of demand. At $5/MMBtu Henry Hub, industrial users (e.g., nitrogen fertilizer producers) curtail operations, reducing demand by 1–2 Bcf/d. At $2/MMBtu, utilities build gas-powered plants and increase firing, while industrial users run at full capacity. This creates a natural price band of roughly $2.00 to $5.00/MMBtu for Henry Hub under normal conditions. The center of this band varies with production costs: at current input costs (steel, labor, diesel), the marginal cost of a new dry gas well in the Haynesville is approximately $2.75/MMBtu. Below $2.50, the market is pricing in “distressed” production; above $4.00, the market is pricing in a demand shock or supply outage. For the 2025 outlook, this band is shifting upward to $2.50–$5.00 due to higher breakeven costs, but it remains the most reliable long-term mean-reversion anchor.

23. The Volatility of the “Thin Market” During Shoulder Months
April, May, September, and October are structurally “thin” months for natural gas trading: lower demand, fewer hedges rolling, and smaller daily storage draws. In these periods, a single storage miss or a unplanned pipeline outage can cause a price move of 5–7% in a single session. Historically, 20% of the annual price return is captured during these four months. For traders, these are the highest alpha periods. The 2025 shoulder months will be particularly sensitive to the LNG export loading schedule: if three LNG tankers are delayed simultaneously, the domestic surplus can grow rapidly; if they accelerate, the surplus disappears. Use weekly vessel-tracking data (e.g., MarineTraffic) to anticipate local basis moves.

24. The Impact of the Clean Air Act and Regional Haze Rules
Utilities in the Midwest and Southeast are facing more stringent regional haze and mercury and air toxics standards (MATS) under the Clean Air Act. Compliance requires installation of SO2 scrubbers and baghouses or switching to lower sulfur fuel—natural gas. The timeline for MATS compliance (EPA deadlines in 2025–2027) will force coal unit retirements (or fuel switching) at roughly 10 GW of capacity, adding 1.0–1.5 Bcf/d of incremental annual gas demand. This regulatory-driven structural shift is a reliable low-volatility bullish factor for the second half of the decade.

25. The Real Estate Development and Housing Starts Link
New home construction in the US directly adds natural gas lines in the majority of new builds (excluding climate zones dominated by heat pumps). The National Association of Home Builders projects 1.4 million housing starts in 2025 (down from 2023 peaks but still elevated). Each new housing unit adds roughly 0.1–0.2 MMBtu/day of peak gas demand. For a 1.4 million-projected housing-start year, this translates to 0.14–0.28 Bcf/d of incremental demand. While small relative to the total market, it is additive and tends to be concentrated in the Sun Belt (Texas, Florida, Arizona), where summer cooling demand is already high. The interplay between mortgage rates, housing affordability, and gas demand is a slow-moving but steady tailwind.

26. The Influence of Temporary Storage Inversion and Arbitrage
In the US, market participants use parking and lending agreements to arbitrage cash versus futures markets. A cash-forward spread exceeding the cost of carry plus 2 cents/MMBtu triggers a storage trade. When the market is in contango (future prices higher), storage operators inject gas to sell later; when backwardated, they push gas out. The amount of working gas in transit (pipeline linepack) also shifts. A large contango structure (e.g., May 2025 futures at $2.50 vs. June 2025 at $2.80) encourages aggressive injection out of May into June, suppressing April/May prices. Tracking the shape of the first six-month spread is crucial for monthly trading.

27. The Intersection of Natural Gas and Bitcoin Mining
Bitcoin mining operations have moved to the US (Texas, New York) and increasingly utilize flared or curtailed natural gas in remote locations. Companies like Crusoe Energy and ExaPower deploy modular data centers that consume 2–5 MW each, burning associated gas that would otherwise be flared. The total demand from this sector is small (estimated 0.3–0.5 Bcf/d in 2025) but rapidly growing and highly price-sensitive: when gas prices fall below $2.00/MMBtu, miners increase utilization; when above $4.00, some shut down. This creates a price-responsive demand buffer that softens downside volatility. For the 2025 outlook, the EIA is starting to track this as a sub-sector in its Monthly Energy Review.

28. The Risk of a Sudden Supply-Push from Venezuela or Iran
Although OPEC+ dynamics are focused on oil, a potential easing of US sanctions on Venezuela (which has significant associated gas reserves) or an Iranian nuclear deal could release 0.3–0.8 Bcf/d of LNG onto the global market. This would reduce the Atlantic Basin’s reliance on US LNG, softening TTF and consequently pulling US Henry Hub lower. While low probability (<10%) in the 2025–2026 window, it remains a tail risk scenario that would collapse international gas spreads. Conversely, a geopolitical shock in the Middle East (Strait of Hormuz) would spike global LNG prices and directly lift Henry Hub, as US cargoes become the marginal swing supply for Europe and Asia.

29. The Data Science of Natural Gas Forecasts: Machine Learning vs. Fundamentals
A growing number of traders and utilities use machine learning models (random forests, LSTM neural networks) to forecast storage and price. These models now outperform simple linear regression on short-term horizons (1–14 days). The dominant features include: temperature ensembles (GFS, ECMWF), pipeline nomination data (Interstate Energy Commission), and storage withdrawal curves. The rise of AI-driven heatmaps (e.g., EIA’s “Real-Time Storage” experimental data) has made the market more efficient in response to weather changes. For the retail or institutional forecaster, the key is not to fight the machine but to identify when machine models fail: during extreme outliers (cold snaps, hurricanes) when fundamental human judgment based on physical market structure outperforms.

30. The Final Factor: The Index Rebalancing Effect
Every month, natural gas futures and swaps are rebalanced by passive index funds (Bloomberg Commodity Index, S&P GSCI). These rebalances create a predictable price impact on the contract month rolling. The “GSCI roll” period (typically five days around the 10th to 14th business day of the month) can move the spread between the front month and second month by 2–4 cents per MMBtu. For active traders, being long the front month during a roll can generate a small positive carry if the market is backwardated. In oversupplied conditions, the roll is negative. Understanding the timing and direction of the roll is essential for portfolio asset managers, but for the long-term outlook, index flows are noise relative to the fundamental factors enumerated above.