Finance Regime Dashboard

Valuation multiples — CAPE & trailing P/E

trailing_pe: 30.4trailing_pe_percentile_since_1871: 97.0cape: 40.1cape_percentile_since_1871: 99.0

CAPE at 100th-percentile since 1871 — today's valuations are at or near historic records (1999-2000 dot-com is the only real comparison). Trailing P/E is lower because current earnings are at cyclical highs (opposite of 2009). Caveat: modern index composition and rate regime may have shifted 'normal' CAPE up from its 150-year mean.

Formula + definitions + sources

CAPE = Price ÷ mean(10 years of inflation-adjusted EPS). Trailing P/E = Price ÷ TTM EPS.

Definitions

P/E (Price-to-Earnings) ratio: Price per share divided by earnings per share. Tells you how many dollars you pay for each dollar of annual profit the company produces. Higher = more expensive.
Trailing 12 Months (TTM): The most recent 4 quarters of reported earnings summed together. A rolling 12-month window that updates each quarter.
Shiller CAPE (Cyclically-Adjusted P/E): Price divided by the AVERAGE of the last 10 years of earnings, adjusted for inflation. Designed to filter out cyclical booms/busts so you're comparing price to a normal-cycle earnings level. Named after Nobel laureate Robert Shiller (2013).
Percentile (since 1871): Where today's reading sits within the FULL Shiller dataset history (1871 → present, 150+ years). 100th percentile = the highest reading ever recorded. 50th = historical median.
Long-run averages: CAPE: 17.3× (median ~16.0×) across full history since 1871. Trailing P/E: 15.8× (median ~14.9×). These are different because CAPE smooths over 10 years of cyclical earnings while trailing P/E responds to current-quarter earnings.
Caveats (per GPT-5.4 sanity check): CAPE is affected by accounting-rule changes (buyback/tax treatment shifts since 2001), sector-composition evolution (asset-light tech now dominates), and ultra-low real rates post-GFC. These shift the 'normal' upward over time — so a reading at 100th percentile vs 1871-2026 is still notable but arguably less dramatic than raw number suggests.

Sources: Robert Shiller's dataset (http://www.econ.yale.edu/~shiller/data.htm) — primary academic source for CAPE, multpl.com (retail-friendly aggregator of Shiller data), For best practice: S&P Dow Jones Indices / FactSet for index earnings; Shiller for CAPE

Refreshed: 2026-04-22T11:08:38.838425

Trailing Earnings-Yield Spread vs 10Y Treasury (ERP proxy)

spread_pct: -0.97spread_percentile_25y: 11.025y_avg_spread_pct: 1.24

Negative trailing-earnings-yield spread (−1%) means the 12-month trailing earnings yield is below the 10Y Treasury yield. Rare but not unique — also happened in 2008-09 (earnings collapse) and 2020 briefly. For a CLEAN read, use forward earnings yield or implied ERP models (Damodaran style) — this proxy is directional only.

Formula + definitions + sources

Spread = (100 / trailing P/E) − 10Y Treasury yield

Definitions

Earnings Yield: Inverse of P/E, expressed as a percentage. If trailing P/E is 30, earnings yield is 1/30 = 3.33%. The fraction of your stock purchase price that comes back as annual trailing earnings (NOT dividends — total earnings, reinvested + paid out).
10-Year Treasury Yield: The interest rate the US government pays on 10-year bonds. Considered the 'risk-free rate' for US dollar investors.
Equity Risk Premium (ERP), proper definition: The EXPECTED extra return investors demand for holding stocks vs risk-free bonds. A TRUE ERP uses forward earnings estimates or implied expectations (Damodaran, Shiller-style models). THIS chart is the backward-looking TRAILING proxy — noisier, cyclical. Useful as directional context but not a precise ERP.
What a negative reading actually means: Trailing earnings yield < 10Y Treasury yield. Said another way: the cash earnings stocks have ACTUALLY produced over the last year are smaller per dollar invested than what bonds pay. Does NOT mean stocks have worse expected future returns — just that current trailing earnings aren't compensating for the bond-yield alternative at current prices.
Historical negative-spread episodes: Also negative during 2008-09 (trailing earnings collapsed in GFC, pushing denominator near zero), 2020 COVID, and 2024-early 2026. Not unique to 2000 dot-com era.

Sources: multpl.com (trailing P/E), FRED series DGS10 (10Y Treasury constant-maturity)

Refreshed: 2026-04-22T11:08:39.062651

S&P 500 TTM Earnings YoY growth (uncapped)

eps_yoy_pct: 13.5eps_yoy_25y_median: 9.1eps_yoy_25y_mean: 25.5

Current +13.5% YoY is ABOVE the 25Y MEDIAN (~+8%) but BELOW the 25Y mean (~+25% — mean is badly skewed by the 2010 base-effect spike from Q1 2009 earnings collapse). Median is the better 'typical' benchmark. Current earnings growth is OK but nothing extraordinary — not supporting today's record valuations (Panel 1a).

Formula + definitions + sources

YoY% = (TTM EPS_t − TTM EPS_{t-12M}) / |TTM EPS_{t-12M}| × 100

Definitions

TTM EPS (Trailing Twelve Months Earnings Per Share): Aggregated earnings of S&P 500 companies over the last 4 quarters, expressed per index share. A lagging indicator reflecting the past year's earnings.
YoY%: Year-over-year percentage change — this month's TTM EPS vs the same month 12 months ago.
2010 spike explained: In Q1 2009, SPX TTM EPS collapsed to $10.65 (banks recognizing massive GFC writedowns). By Q1 2010, TTM EPS recovered to ~$62. That's +480% YoY — a real number showing base-effect, not a data artifact. Previously capped at +250% for readability; now shown uncapped per data-honesty principle.

Sources: multpl.com (S&P 500 TTM EPS), Robert Shiller's Yale dataset (primary)

Refreshed: 2026-04-22T11:08:39.353923

VIX — Implied Volatility (fear gauge)

vix_level: 18.87percentile_since_2001: 58.025y_avg_approx: 19.7

Today VIX at 18.9 (58th percentile since 2001). Classic 'calm market, high valuations' setup when combined with Panel 1a's record CAPE — complacency alongside elevated prices has historically preceded volatility expansions.

Formula + definitions + sources

VIX = Chicago Board Options Exchange Volatility Index — 30-day annualized implied volatility derived from S&P 500 option prices. Daily close.

Definitions

VIX / Volatility Index: A real-time forward-looking measure of expected stock market volatility over the next 30 days, calculated by the CBOE from prices of S&P 500 put and call options. Expressed as an annualized percentage.
Implied volatility: The level of volatility the OPTIONS MARKET is pricing in — not what happened historically (that's 'realized volatility'), but what traders expect. Inferred by inverting the Black-Scholes formula from observed option prices.
Why 'fear gauge': When traders expect big moves (especially downside), they pay more for options (put protection, call upside) → option prices rise → implied volatility rises → VIX rises. Low VIX = complacent market. High VIX = priced-in anxiety.
Regime bands: Under 15 = calm / complacent / risk-on. 15–25 = normal / elevated. 25–40 = stressed. >40 = crisis-level (2008 GFC, 2020 COVID, 2018 volmageddon).
Relationship to valuation: VIX is the market's FORWARD-looking risk assessment; the P/E ratios in Panel 1a are BACKWARD-looking valuation. When VIX is low AND valuations are high, the market is pricing 'smooth sailing at record levels' — historically a complacency signal, not a robust bullish read.

Sources: FRED VIXCLS (CBOE, daily close)

Refreshed: 2026-04-22T11:33:04.848157

Earnings by cohort

Earnings by Cohort — TTM Level ($B)

mag6_ttm_b: 627.0tsla_ttm_b: 4.0rest_ttm_b: 1365.0

Dollar trajectory is the cleanest view — no ratio artifacts. Mag6 clearly dominant; Tesla visibly shrinking; Rest-of-SPX growing steadily.

Formula + definitions + sources

TTM NI = sum of most recent 4 quarterly net income values at each quarter end. Mag7-ex-Tesla TTM = sum of 6 tickers' TTMs (AAPL, MSFT, GOOGL, AMZN, META, NVDA). SPX TTM total = (SPX MC ÷ SPX PE) — derived via aggregate identity, not summed. Rest TTM = SPX TTM − Mag6 TTM − Tesla TTM.

Sources: macrotrends.net (Mag6 + Tesla cash-flow statements, SEC 10-Q/10-K), Siblis Research (SPX year-end market cap), multpl.com (SPX trailing P/E, Shiller dataset), Yahoo Finance ^GSPC (intra-year price scaling)

Refreshed: 2026-04-22T11:08:45.230309

Earnings by Cohort — Raw Quarterly NI ($B) with YoY %

mag6_latest_q_b: 202.0tsla_latest_q_b: 0.9

Raw quarterly catches inflection moments better than TTM (no 4-quarter smoothing). YoY % compares each quarter to same quarter a year ago — filters seasonality, shows real annual pace.

Formula + definitions + sources

Mag6 quarterly NI = sum of 6 tickers' reported quarterly GAAP NI. Tesla = TSLA standalone GAAP NI. Rest = SPX total TTM ÷ 4 (proxy; can't cleanly decompose raw quarterly). YoY % labels = (current quarter − same quarter 1 year ago) / |that prior quarter| × 100.

Sources: macrotrends.net cash-flow statements, Siblis Research + multpl.com + Yahoo Finance (SPX)

Refreshed: 2026-04-22T11:08:45.351314

CapEx

Mag7 CapEx — Raw Quarterly ($B)

mag7_q4_2025_capex_b: 123.6qoq_pct: 19.0yoy_pct: 60.0

Raw quarterly view catches inflection moments. Dominant contributors visible as stacked bands — AMZN/GOOGL/MSFT/META are the quartet driving the AI-era capex acceleration.

Formula + definitions + sources

CapEx per ticker per quarter = abs('Net Change in Property, Plant & Equipment') from quarterly cash flow. Mag7 quarterly total = sum across 7 tickers for a given calendar quarter.

Sources: macrotrends.net per-ticker cash-flow-statement (ultimately SEC 10-Q/10-K filings)

Refreshed: 2026-04-22T11:08:39.610688

Mag7 capex vs total US business capex

mag7_ttm_b: 399.0pnfi_ttm_b: 4251.0mag7_share_of_us_pct: 9.4mag7_share_2011_pct: 0.9

Mag7 represents 9.4% of all US private non-residential capex — up from 0.9% in 2011. ~10x concentration increase.

Formula + definitions + sources

Mag7 TTM = rolling 4Q sum. PNFI TTM = mean of last 4 SAAR values (actual TTM annualized). Share = Mag7 TTM / PNFI TTM × 100.

Sources: macrotrends.net (Mag7), FRED series PNFI (Bureau of Economic Analysis)

Refreshed: 2026-04-22T11:08:39.819727

Mag7 CapEx ROI proxy (1Y and 2Y payback denominators)

roi_1y_pct: 35.0roi_2y_pct: 22.0roi_1y_hist_avg: 34.0roi_2y_hist_avg: 19.0

Both current readings are ROUGHLY AT OR SLIGHTLY ABOVE their historical averages in this (post-2016) window, not below — the 'spending ahead of payback' concern is real but hasn't yet shown up as ROI underperformance. Real test is whether 2026-2028 earnings acceleration keeps this above historical as capex denominators keep ballooning.

Formula + definitions + sources

Numerator: ΔTTM_NI over 1 year. Denominator: cumulative CapEx over 4 quarters (1Y) or 8 quarters (2Y). Crude — mixes payback from older capex with ongoing recent investment; real ROI emerges 18-30 months post-deployment.

Sources: macrotrends.net quarterly cash-flow statements (SEC 10-Q/10-K)

Refreshed: 2026-04-22T11:08:39.960914

Macro

Inflation & Monetary Policy

cpi_yoy_pct: 3.29fed_funds_pct: 3.64

Gap between Fed Funds and CPI is a shortcut for real policy rate — positive means restrictive. Note Fed's actual 2% target is PCE, not CPI.

Formula + definitions + sources

CPI YoY = (CPI_t − CPI_{t-12M}) / CPI_{t-12M} × 100. Fed Funds = effective daily rate, last monthly value.

Definitions

CPI (Consumer Price Index): Measures the average change over time in prices paid by urban consumers for a market basket of consumer goods and services. BLS-published. The CPI YoY% shown here is the most common inflation headline.
Fed Funds Rate: The interest rate banks charge each other for overnight loans of reserves. The Federal Reserve's primary tool for influencing short-term rates. 'Effective' = actual market-clearing rate, not the Fed's target range midpoint.
Fed's inflation target: The Fed targets 2% inflation as measured by the PCE (Personal Consumption Expenditures) deflator, NOT the CPI shown here. PCE runs typically ~0.3-0.4pp lower than CPI due to methodology differences (housing weight, basket adjustments). The 2% dotted line on chart is for context; strict comparison would need PCE.
Real interest rate (rough shortcut): Fed Funds − CPI YoY ≈ the inflation-adjusted policy rate. Positive = restrictive policy (tightening financial conditions). Negative = stimulative. A more precise calc would use EXPECTED inflation (not trailing), commonly from the 5Y5Y forward inflation swap or Cleveland Fed's expectations series.

Sources: FRED CPIAUCSL (BLS Consumer Price Index, seasonally adjusted), FRED DFF (Federal Reserve effective federal funds rate)

Refreshed: 2026-04-22T11:08:40.256529

Labor Market — Unemployment & Job Openings

unrate_pct: 4.3jolts_openings_m: 6.9

Today JOLTS 6.9M openings vs ~7.2M unemployed — roughly at equilibrium. 2022 peak was 12M openings vs 5.6M unemployed (very tight labor market). Recent softening is meaningful — employers slowing hiring, unemployment ticking up. Watch for JOLTS to fall BEFORE unemployment rises sharply — that's the classic recession setup.

Formula + definitions + sources

Unemployment Rate = unemployed people / labor force × 100. JOLTS Openings = total non-farm job postings (count of open positions).

Definitions

Unemployment Rate: % of the labor force that is jobless and actively seeking work. BLS Household Survey. Doesn't include people who've given up looking ('discouraged workers') — just active job-seekers.
JOLTS (Job Openings and Labor Turnover Survey): BLS establishment survey tracking how many jobs EMPLOYERS have open. Counts actual vacancies where employers are actively hiring.
Labor force: Everyone employed PLUS everyone actively seeking work. ~170M people in the US.
'Tight' vs 'loose' labor market: Counterintuitive wording! 'TIGHT' means labor SUPPLY is tight/scarce — MORE jobs than job-seekers, employers compete for workers, wages rise. Good for workers. 'LOOSE' means labor is abundant — more seekers than jobs, employers set the terms, wages stagnate. Good for employers. The word 'tight' is from the employer's perspective (scarce labor supply), not the worker's.
Why this matters: When JOLTS openings COLLAPSE while unemployment is still low, that's the classic recession lead-indicator — employers stop hiring BEFORE they start firing. Watching the openings-to-unemployed ratio is one of the cleanest labor-cycle signals.

Sources: FRED UNRATE (BLS Household Survey), FRED JTSJOL (JOLTS, BLS establishment survey)

Refreshed: 2026-04-22T11:08:40.501153

Real GDP growth (BEA headline QoQ SAAR + YoY context)

real_gdp_qoq_saar_pct: 0.5real_gdp_yoy_pct: 1.9925y_avg_saar_pct: 2.25

Q4 2025 SAAR dropped to 0.5% — economic growth nearly stalled for the quarter. YoY at ~2% suggests the slowdown is recent; trailing year still positive. If Q1 2026 is also weak, recession watch begins.

Formula + definitions + sources

QoQ SAAR = percentage change in real GDP from prior quarter, annualized (BEA's headline metric). YoY = (Real GDP_t − Real GDP_{t-4Q}) / Real GDP_{t-4Q} × 100.

Definitions

Real GDP: Gross Domestic Product adjusted for inflation. Measured in 'chained 2017 dollars' — a base-year pricing convention that strips out general price level changes so you see REAL output growth.
SAAR (Seasonally Adjusted Annual Rate): A quarter's growth rate, annualized. If GDP grows 1% in one quarter, SAAR is ~4.1% (1.01^4 − 1). This is what the BEA and financial media report as 'GDP growth' — it's the POINT-IN-TIME velocity.
QoQ vs YoY: QoQ SAAR tells you the economy's current pace — volatile, sensitive to this quarter's data. YoY tells you the trailing-year trend — smoother, less sensitive to any single quarter. Both are in the chart so you can see 'current velocity' and 'annual trend' together.
Recessions: Classically defined as 2 consecutive negative QoQ SAAR quarters. Q1 2025 was negative (−0.6%) but Q2-Q3 rebounded strongly, so no recession declared.

Sources: FRED A191RL1Q225SBEA (BEA headline QoQ SAAR), FRED GDPC1 (BEA Real GDP level)

Refreshed: 2026-04-22T11:08:40.750624

Credit / Leverage

High-Yield Credit Spread (HY OAS)

hy_spread_pct: 2.87hy_25y_avg: 3.32hy_25y_percentile: 22.0

HY spreads widen before equity drawdowns. Anything above 4% is stress territory. Current at 2.87% = below average, no distress being priced in by credit markets.

Formula + definitions + sources

OAS of ICE BofA US HY index = bond yield of HY basket − yield of matched-duration US Treasury, adjusted for embedded-option values.

Definitions

High Yield (HY) bonds: Corporate bonds rated below investment grade (BB+ and below at S&P; Ba1 and below at Moody's). Also called 'junk bonds'. Higher default risk than investment-grade; issuers pay higher interest rates to compensate.
Spread: The EXTRA yield a HY bond pays over a US Treasury bond of similar maturity. If a 5-year Treasury yields 4% and a 5-year HY bond yields 7%, the spread is 3 percentage points (300 basis points).
Option-Adjusted (OAS): Corporate bonds often have embedded options (callable = issuer can redeem early; putable = investor can demand early repayment). These options have value. OAS strips out that value so the spread reflects ONLY credit risk, not option dynamics.
ICE BofA US High Yield Master II Index: The standard HY bond benchmark. ~2000 bonds from US corporate issuers rated below investment grade. Index tracks aggregate yield, duration, credit quality.
Why it matters: HY spreads are a credit-market stress gauge. They widen during risk-off episodes and narrow when market is complacent. Historical reference: GFC 2008 hit 20%, COVID 2020 hit 11%, 2022 rate scare ~5%. >4% is meaningful stress. The relationship with equity drawdowns is context-dependent — sometimes spreads lead, sometimes lag, and sometimes they move together; they're best read as a confirmatory signal rather than a reliable leading indicator.

Sources: FRED BAMLH0A0HYM2 (ICE BofA)

Refreshed: 2026-04-22T11:08:40.849025

Corporate Debt / GDP

debt_to_gdp_pct: 45.1percentile_25y: 42.0

Rising corporate debt/GDP with flat/falling GDP growth = leverage risk. 2020 COVID spike was borrowing-through-the-crisis.

Formula + definitions + sources

Non-financial corporate debt securities ÷ nominal US GDP × 100. Approximates aggregate corporate leverage.

Sources: FRED NCBEILQ027S (Flow of Funds), FRED GDP (BEA)

Refreshed: 2026-04-22T11:08:41.112898

Mag7 Leverage — Net Debt / EBITDA snapshot

TSLA: {'ratio': -3.78, 'net_debt_b': -35.9}GOOGL: {'ratio': -0.96, 'net_debt_b': -144.3}MSFT: {'ratio': -0.43, 'net_debt_b': -69.0}META: {'ratio': -0.35, 'net_debt_b': -36.0}AMZN: {'ratio': -0.34, 'net_debt_b': -50.1}NVDA: {'ratio': -0.18, 'net_debt_b': -23.4}AAPL: {'ratio': -0.14, 'net_debt_b': -20.9}

Negative ratios (green) = net cash: company has more cash/investments than debt. 0–1.5× (amber) = low leverage. >2× (rose) = meaningful leverage. Mag7 is almost entirely net-cash — one of the few cohorts where this is true even during a capex boom.

Formula + definitions + sources

Net Debt = Total Debt − (Cash + Current Investments + Non-Current Investments). EBITDA = Operating Income + Depreciation & Amortization. Ratio = Net Debt ÷ EBITDA.

Definitions

Net Debt: Total gross debt MINUS all liquid assets (cash + short-term investments like Treasuries + long-term investments). Answers 'how much would the company owe if they liquidated all investments to pay debt?' Negative = they'd have money left over (net cash position).
EBITDA: Earnings Before Interest, Taxes, Depreciation, and Amortization. A proxy for 'cash operating earnings' — pre-financing, pre-tax, pre-non-cash-charges. Used because it approximates the cash flow available to service debt.
Net Debt / EBITDA: How many YEARS of EBITDA would it take to pay off net debt. Lower = safer leverage. Common rules of thumb: under 1× = very safe. 1–3× = moderate. Above 4× = stressed. Negative (net cash) = fortress balance sheet.
Why this matters: In a rate-rising cycle, highly-leveraged companies face squeeze (interest costs climb; debt service chews cash). Net-cash companies are immune — they earn INTEREST on their cash pile. This is why Mag7 has had the cash pile that's funding the AI capex buildout without borrowing.

Sources: Financial Datasets API (financialdatasets.ai) — SEC 10-K annual filings

Refreshed: 2026-04-22T11:08:45.089294

Energy

Energy Spot Prices — WTI + Henry Hub

wti_usd_per_bbl: 100.72henry_hub_usd_per_mmbtu: 2.79latest_date: 2026-04-13

Today: WTI around $100 (elevated), Henry Hub around $2.80 (low). The divergence (oil high / gas low) is typical of the current US energy regime — domestic gas abundance from shale, oil more tightly coupled to global markets.

Formula + definitions + sources

Daily spot prices as reported by the EIA.

Definitions

WTI (West Texas Intermediate): US benchmark crude oil grade, light and sweet (low sulfur). Priced at the Cushing, Oklahoma pipeline hub. 'Spot' = immediate-delivery cash market price, $/barrel.
Henry Hub: The pricing point for North American natural gas futures and physical trade — a pipeline junction in Erath, Louisiana. The US benchmark for gas prices, $/MMBtu (million British thermal units).
Why benchmark prices matter: These are the CORE inputs to inflation, corporate margins, household heating/transport costs, and (currently most relevant) the electricity generation cost stack that powers AI datacenters. Rising fuel costs can compress margins across the economy, rising power costs specifically pressure tech-stock marginal economics.
MMBtu vs $/bbl: Different units because oil and gas are sold by different denominations. MMBtu (heat content) lets you compare across fuels for power-generation economics. Crude WTI is typically 10-20× Henry Hub NG on per-MMBtu basis.

Sources: US EIA API v2 — petroleum/pri/spt (WTI), US EIA API v2 — natural-gas/pri/fut (Henry Hub)

Refreshed: 2026-04-22T12:00:20.972979

US Electricity — Retail Price (monthly) + Generation (12M rolling)

retail_price_cents_per_kwh: 13.69latest_price_date: 2026-01-01generation_ttm_twh_annualized: 4427.0latest_gen_date: 2026-01-01

Retail price 13.7¢/kWh, US total electricity generation running at 4427 TWh/year (12-month rolling). Watch the price line for inflation-pass-through; watch generation-TTM for AI-datacenter capacity expansion (real signal, not seasonal noise).

Formula + definitions + sources

Retail price = national average across all sectors (residential + commercial + industrial), monthly, cents/kWh. Generation = trailing 12-month SUM of all US electricity generation, TWh/year (annualized). Rolling sum strips out seasonal swings (summer AC peaks, spring/fall troughs).

Definitions

Retail electricity price: The average $/kWh (expressed in cents) that US consumers actually pay on their electricity bills — weighted across residential, commercial, industrial customers. Includes fuel costs, generation markup, transmission, distribution, and utility profit.
Wholesale vs retail: Retail is the end-user price (your power bill). Wholesale is the price generators sell at — typically 30-50% lower (no distribution/T&D markup). For macro-regime purposes the retail price captures pass-through to economy.
Generation (TWh): How much electricity the US actually produced that month. 1 TWh = 1,000,000 MWh = 1,000,000,000 kWh. Reflects combined output from coal, gas, nuclear, renewables, and hydro.
Why this matters for equities: Electricity IS the critical input for AI datacenter economics. Rising retail/wholesale prices squeeze datacenter margins and capex ROI; rising generation capacity expands what's possible. Watch these together — a gap between rising demand (gen) and rising price (inflation-indexing) tells the story.

Sources: US EIA API v2 — electricity/retail-sales, US EIA API v2 — electricity/electric-power-operational-data

Refreshed: 2026-04-22T12:00:21.236845