Glossary: terms starting with E

A crypto-asset under MiCA that aims to maintain a stable value by referencing a single official currency (EMT). Only credit institutions and e-money institutions may issue it; holders can redeem it at par at any time, and paying interest is prohibited. At least 30% of the funds received are held in segregated accounts with credit institutions.

Example: $500 million in circulation: at least 0.3 × 500 = $150 million as deposits with credit institutions, the remainder of up to $350 million in secure, highly liquid assets in the same currency.

Source: MiCA (2023); BaFin, MiCAR (2024)

More: Digital assets: tokenization, stablecoins, crypto securities and the digital euro

Inverse of the P/E: earnings per share divided by the price. A P/E of 25 corresponds to a 4% earnings yield; comparing it with bond yields is common but imprecise, because earnings grow in real terms.

More: Shiller P/E (CAPE): how expensive is the stock market?

A durable competitive advantage that keeps competitors away and protects high returns on capital for years, e.g. network effects, switching costs, cost advantages of scale, or protected brands, patents and licenses.

More: Fundamental analysis: valuing a company like an investor

Electronic Data Gathering, Analysis, and Retrieval: the SEC's electronic filing and retrieval system. Company annual and quarterly reports, insider filings and public 13F filings all go through it, and all of them can be searched free of charge. For the holdings of well-known investors, EDGAR is the primary source.

Source: SEC, Form 13F (2026)

More: Legendary investors: the investment styles of the most famous investors, from Graham to Simons

A measure of how many uncorrelated, equally sized risks a portfolio takes: the exponential of the entropy of the variance shares that uncorrelated building blocks (principal components or minimum-torsion factors) contribute to portfolio risk. 1 means a single risk, N means N equally sized ones.

Example: Variance shares of 50% and 50%: exp(−2 × 0.5 × ln 0.5) = 2 bets; shares of 90% and 10%: 1.38 bets.

Source: Meucci (2009)

More: Beyond Markowitz: estimation error, noise and true diversification · Methodology: Dimensions methodology: how Portfolio Lab counts independent risks

1 divided by the sum of the squared weights: how many equally sized positions would give the same diversification?

Example: Weights of 40, 30, 20 and 10% give 3.33.

More: Look-through: what is really inside your ETFs

The number of truly independent variants tried in a strategy search. Similar variants, such as neighboring parameters, are highly correlated and together count for less than their number; López de Prado and Lewis (2019) estimate the effective number with a clustering method. It is the N that goes into the deflated Sharpe ratio.

Source: López de Prado and Lewis (2019); Bailey and López de Prado (2014)

More: Backtest overfitting: when the past looks too good

All portfolios that offer the highest expected return for their level of risk (Markowitz, 1952).

More: Efficient frontier: Markowitz portfolio optimization

The thesis that prices reflect all available information, so that this information cannot be used to earn systematic excess returns. Fama distinguished a weak, a semi-strong and a strong form.

Source: Fama (1970); Malkiel (2003)

More: The efficient market hypothesis: why it is so hard to beat the market

In principal component analysis, the variance explained by a principal component. The share of the largest eigenvalue in the total shows how strongly a single factor drives all assets together.

Source: Pearson (1901)

Methodology: Asset analysis: methodology from correlation to cointegration

European Long-Term Investment Fund: an EU fund type for unlisted companies, loans, infrastructure and real estate. Since ELTIF 2.0 (from January 10, 2024), at least 55% must be held in eligible assets; retail investors may buy after a suitability assessment and without a statutory minimum investment.

Source: Regulation (EU) 2023/606 (2023)

Expectation-maximization: a method for maximum likelihood estimation with missing or hidden data. It alternates between the E-step, which uses the current parameters to estimate the hidden variables as probabilities, and the M-step, which re-estimates the parameters with these weights. Typical applications are mixture distributions such as the Gaussian mixture model, and hidden Markov models.

Source: Dempster et al. (1977)

More: Regime models: hidden Markov, Markov switching and more · Methodology: Regime methodology: how the hidden Markov model works

A middle-income country whose capital market is not yet fully developed, e.g. China, India or Brazil. Higher volatility, with its own political and currency risks.

More: World portfolio: the whole world in one or two ETFs

Bond issued by an emerging market country, either in hard currency (usually US dollars) or in local currency. With local currency, exchange rate risk comes on top of credit quality.

Benefit that owning a real asset provides in itself: pleasure in a painting, the joy of driving, prestige. It replaces part of the financial return: according to Mandel (2009), this is why the financial returns of art are low; Dimson et al. (2015) find a growing non-financial benefit in mature fine wines.

Source: Mandel (2009); Dimson et al. (2015)

More: Collectibles as an investment: what art, wine, watches and classic cars really return

Shares of the company you work for, e.g. from employee share schemes. Income and wealth then depend on the same risk; according to studies, the lack of diversification costs a large part of the market value.

Source: Benartzi (2001); Meulbroek (2005)

More: Your whole wealth: total wealth from portfolio, home, human capital and car

Investment approach of large university endowments, shaped by David Swensen at Yale: broad diversification across many asset classes, few bonds and a high share of illiquid assets such as private equity, real estate and hedge funds, to profit from illiquidity premiums and good managers. Requires a very long investment horizon.

Source: Swensen (2000)

More: Legendary investors: the investment styles of the most famous investors, from Graham to Simons

A two-step test for cointegration: first a regression estimates the relationship between two series, then a unit root test checks whether the residuals are stationary. Because the relationship is estimated, stricter critical values apply; MacKinnon's formulas provide the p-values.

Source: Engle and Granger (1987); MacKinnon (1994)

More: Pairs trading and statistical arbitrage: idea, methods, research · Methodology: Long/short and statistical arbitrage: methodology

Euro OverNight Index Average, the former reference rate for overnight money in the euro area until 2021. For long time series it is chained with the €STR.

Methodology: Metrics methodology: how Portfolio Lab measures return and risk

Every asset receives the same share (1/N). In out-of-sample studies this simple rule beats many optimization models because it picks up no estimation error.

Source: DeMiguel et al. (2009)

More: Risk parity: equal risk contributions instead of equal weights

An index in which every stock has the same weight. It tilts toward smaller companies and requires regular rebalancing back to equal weights.

The puzzle that the historical equity premium was far higher than standard economic models with reasonable risk aversion can explain. Proposed explanations range from loss aversion to rare disasters.

Source: Mehra and Prescott (1985); Benartzi and Thaler (1995)

More: Equity risk premium: how much more equities return and why it is a puzzle

The excess return of equities over safe assets as compensation for their risk. Historically, in many countries, it amounted to several percentage points a year over decades, with large differences between countries and periods.

Source: Dimson et al. (2002); Mehra and Prescott (1985)

More: Equity risk premium: how much more equities return and why it is a puzzle

A criticism of mean-variance optimization: it favors precisely those assets whose return has been overestimated and whose risk has been underestimated. As a result, the optimized weights are often extreme and unstable.

Source: Michaud (1989); Chopra and Ziemba (1993)

More: Efficient frontier: Markowitz portfolio optimization

Short for environmental, social and governance, i.e. the environment, social issues and good corporate governance. An umbrella term for the criteria investors use to assess companies beyond their financial metrics.

More: ESG ETFs: index rules, ratings, returns and the limits of sustainable funds

A rating provider's score for how well a company meets ESG criteria. Ratings from different providers diverge widely, mainly because they measure the same categories differently and include different categories.

Source: Berg et al. (2022)

More: ESG ETFs: index rules, ratings, returns and the limits of sustainable funds

The gap between a quantity estimated from data, such as an expected return, and its true value. For average returns it shrinks only with the square root of the number of years, and it often drives the weights of an optimization more than the information does.

Source: Michaud (1989); Chopra and Ziemba (1993)

More: Michaud resampled optimization: more stable portfolio weights through simulation

Exchange-traded commodity: an exchange-traded debt security on a commodity, often physically backed, as with gold. It is not a segregated fund: there is issuer risk, which the collateral reduces.

Exchange-traded fund: an investment fund traded on a stock exchange that usually tracks an index. It can be bought and sold like a share during trading hours and is protected as segregated assets.

Source: Ben-David et al. (2017)

More: ETF basics: costs, replication and distributions

Exchange-traded note: an unsecured debt security whose repayment depends on an index. If the issuer defaults, the money is gone, even if the index has risen.

The mathematical reason why risk contributions add up exactly: volatility is homogeneous of degree one in the weights, so it equals the sum of each weight times its partial derivative.

Source: Roncalli (2013)

More: Risk contribution: breaking portfolio risk down into its parts

Enterprise value (market capitalization plus net financial debt) divided by earnings before interest, taxes, depreciation and amortization. Unlike the P/E ratio, this metric hardly depends on how a company is financed; however, it easily makes companies with heavy capital expenditure look cheaper than they are.

More: Fundamental analysis: valuing a company like an investor

A contract whose payout depends on whether an event occurs, usually 1 for yes and 0 for no. Economically, it is a digital option on an event. In the US the CFTC treats such contracts as swaps, which may only be offered on registered trading venues.

Source: CFTC (2022)

More: Prediction markets as bets on events: what they can and cannot do

A method that measures the price reaction to an event such as an earnings release, a takeover or an index inclusion: the normal return is estimated from an estimation window before the event, and the abnormal return is measured in the event window and tested across many events.

Source: MacKinlay (1997); Fama et al. (1969)

More: Event studies: how to measure and test price reactions to events

The first trading day on which a share trades without entitlement to the declared dividend. The price then falls by roughly the amount of the dividend; exactly how much also depends on investors' tax rates.

Source: Elton and Gruber (1970)

Annual, individual statement of all costs and charges actually incurred under MiFID II, as an amount in euros and as a percentage, with an illustration of how the costs reduce the return cumulatively.

Source: European Commission (2017)

More: Portfolio reporting: how to report return, risk, attribution and costs properly

Earnings yield of the equity market (1 / CAPE) minus the real yield on ten-year government bonds: the extra yield of equities over bonds.

More: Shiller P/E (CAPE): how expensive is the stock market?

A concept from stochastic portfolio theory: the amount by which a portfolio's growth rate exceeds the weighted average of its components' growth rates, per period log(Σ wᵢ × (1 + Rᵢ)) − Σ wᵢ × log(1 + Rᵢ). With positive weights it is never negative by Jensen's inequality and is approximately half the weighted dispersion of the log returns around their mean; with short sales or leverage at portfolio level this does not hold (Qian 2012).

Example: 60% equities at +10% and 40% bonds at −5%: log(1.04) − (0.6 × 9.531% + 0.4 × (−5.129%)) = 3.922% − 3.667% = 0.255%.

Source: Fernholz and Shay (1982); Fernholz (2002); Qian (2012)

In stochastic portfolio theory, the gap between a portfolio's growth rate and the weighted average of its positions' growth rates: ½ × (Σ wᵢσᵢ² − σ_p²). Never negative with positive weights, and the less the positions are correlated, the larger it is.

Example: Two uncorrelated assets, each with 20% volatility, 50/50: ½ × (0.04 − 0.02) = 0.01, i.e. 1 percentage point of growth rate a year.

Source: Fernholz (2002)

More: Beyond Markowitz: estimation error, noise and true diversification · Methodology: Dimensions methodology: how Portfolio Lab counts independent risks

The return above a reference point, usually the risk-free rate. Measured against an index, it is called active return.

More: Sharpe ratio: return per unit of risk

Robert Shiller's finding: share prices fluctuate far more than the dividends actually paid later would justify. An argument against strict market efficiency.

Source: Shiller (1980)

More: The efficient market hypothesis: why it is so hard to beat the market

Rules under which a fund does not buy certain companies, e.g. makers of controversial weapons, tobacco companies, coal companies or firms with serious violations of the UN Global Compact.

More: ESG ETFs: index rules, ratings, returns and the limits of sustainable funds

Assumption that long-term interest rates merely reflect the average of expected short-term rates, so that forward rates are exactly the expected rates. It does not hold in the data: forward rates contain term premiums that vary over time.

Source: Campbell and Shiller (1991); Fama (1984)

More: Market expectations: what the yield curve, inflation-linked bonds and options reveal about the future

Expected positive market value of the transactions with a counterparty at a future point in time, i.e. the amount that would be at stake on average in a default. Potential future exposure (PFE), by contrast, is a high quantile of this value, such as 95%.

Source: Gregory (2020)

More: Counterparty and issuer risk: CVA, ETCs, certificates and swaps

The Sharpe ratio that the best of N independent variants with no real edge can be expected to reach by chance alone: SR₀ = dispersion × ((1 − γ) × Φ⁻¹(1 − 1/N) + γ × Φ⁻¹(1 − 1/(N × e))), with the Euler-Mascheroni constant γ. It is the hurdle of the deflated Sharpe ratio and rises with the number and the dispersion of the variants.

Example: Ten variants, dispersion of their Sharpe ratios 0.3 p.a.: 0.3 × (0.4228 × 1.282 + 0.5772 × 1.789) = 0.47 p.a.

Source: Bailey and López de Prado (2014)

More: Deflated Sharpe ratio: how much skill is in a Sharpe ratio?

Probability-weighted average of all possible outcomes; in a simulation, the mean across all paths. For skewed distributions such as final wealth after many years, it is above the median: a few very good paths pull it up.

Example: Lognormal distribution with median 1 and σ = 0.5: expected value e^(0.5² / 2) = 1.13.

More: Volatility drag: why volatility costs you return · Methodology: Methodology of the free calculators: FIRE, savings plan, withdrawal and leveraged ETF

Day on which an option or a future ends. VIX futures expire on a Wednesday 30 days before the third Friday of the following month, the expiry of the corresponding S&P 500 options.

Methodology: Volatility dashboard: methodology

A price that for a time grows faster than a random walk allows. Tests such as the PSY test by Phillips, Shi and Yu check for this with right-tailed unit root tests over changing windows and date the start and end of such phases.

Source: Phillips, Wu and Yu (2011); Phillips, Shi and Yu (2015)

More: Speculative bubbles: what they are and how the LPPLS model detects them

Weight of a segment according to its economic exposure rather than its market value: futures and swaps at their notional value, options at delta times the underlying, plus a financing line so that the weights add up to 100%.

Example: 50% equities plus equity futures worth 10% of assets: exposure weight of equities 50% + 10% = 60%, financing −10%.

More: Multi-asset attribution: weighting schemes, top-down and bottom-up, currencies and overlays

A certificate that is redeemed early with a coupon if the underlying is above a threshold on an observation date. Otherwise it continues; at maturity a barrier protects you, and below it you bear the price loss. Decomposed: a bond plus digital calls minus a down-and-in put.

The statistics of rare, very large events. It models only the tails of a distribution and uses them to estimate losses beyond anything ever observed in short data series.