Financial Education

Previous studies based on returns-based analysis have found no evidence of market-timing ability by mutual funds. In the December 2007 Journal of Financial Economics, Jiang Yao and Yu conduct a holdings-based analysis of 2,300 equity mutual funds and conclude that mutual fund managers have positive and statistically significant market timing ability for three- and six-month periods.

Mutual fund characteristics associated with positive market timing ability include high industry concentrations, large size, and small-capitalization orientations. The authors also find that stronger market timing results are associated with shifting between industries than by adjusting allocations within an industry.

When mutual funds experience large investor outflows, they must often quickly reduce their holdings to return funds to investors. Such forced liquidation can lead to asset sales at “fire sale” prices. In the November 2007 Journal of Financial Economics, Coval and Stafford show that mutual funds do not allow for the risk of such events, and that such flows are predictable – resulting in an incentive to front run the funds.

Funds in the highest and lowest deciles ranked by performance and prior outflows predictably face large inflows and outflows in subsequent periods. Investors can anticipate such flows by buying or selling the largest fund holdings ahead of the cash flows and reversing the position afterward.

The value premium refers to the fact that stocks with priced at a low multiple of book value tend to perform better than would be explained by the Capital Asset Pricing Model. Proponents of the Efficient Market Hypothesis (EMH) argue that the value premium results from a risk factor not captured by Beta. Opponents argue that the value premium is evidence that the EMH is not valid.

Fama and French, among others, have developed models that attribute the value premium to risk factors. Typically these studies use company size as a proxy for liquidity risk. In the Review of Finance, Phalippou replicates these studies in their original form, and then using institutional ownership as an alternative liquidity proxy. In this second iteration, the value premium persists and cannot be explained by the risk factor models.

Phalippou concludes that if the value premium is attributable to risk factors, the risk factors are not the ones identified in previous research.

Brokers and dealers play different roles in securities markets.

The broker is an agent of the investor. He represents the order, finding opposite sides of the trade. Brokers also supply market information, provide discretion and secrecy, and sometimes provide other supporting services such as margin, record-keeping, custody, etc.)

Dealers are adversaries of the investor. They benefit from a higher bid/ask spread, where the investor covets a lower one. Since their profits are small on a given trade, they are unwilling to trade against informed investors (those who have specific information that the security is mispriced) or those with a reputation for finding mispriced securities. The informed investors want to maintain their ability to trade, and thus don’t want dealers to know they are behind the order.

The value premium refers to the well-documented outperformance of value stocks (those with high book value relative to market value) over gorwth stocks (those with high market-book ratios.) This outperformance is not explained by systematic Beta as defined in the CAPM, though it does represent one of the three/four factors used in the Fama-French model.

In the March/April 2008 Financial Analysts Journal, Phalippou finds that the value premium is driven by stocks with low institutional ownership – a group that represents just 7% of the total stock market capitalization. Since individual investors may be less sophisticated than institutional investors, such stocks may have a higher tendency to be mispriced. The low institutional ownership may also signal that the mispricing opportunities are difficult to arbitrage.

A mispricing occurs when the price of a security predictably deviates from its normal or expected return. Persistent mispricings are called anomalies. However, the methods for determining expected return, and of finding anomalies, can introduce bias.

Expected Return

In order to constitute a mispricing, a security must predictably earn a higher or lower return than expected. However, the expected return itself is subjective. If it is improperly measured, an apparent anomaly may be nothing of the sort.

Data Mining

Studying hundreds or thousands of relationships is likely to result in a few that appear significant only due to random chance.

Survivorship Bias

When results are based on existing entities, they may ignore entities that have failed. While the existing fund managers, on average, have outperformed their benchmarks this is because if they had not investors would have withdrawn their funds. Only an examination of all managers, whether failures or successes, can give a true reading.

Small Sample Bias

Patterns observed over a short time period may not repeat in other time periods.

Selection Bias

Some anomalies are affected by a portion of the sample. For example, a certain anomaly may pertain only to small cap stocks. Attempts to exploit the anomaly could fail if not applied to the correct sample.

Nonsynchronous Trading

Stocks that trade less frequently will have price changes that reflect all information since the prior trade. Large swings in the overall market between trades will be masked. The stock may appear less volatile than it actually is.

Risk

Riskier investments should generate higher returns. If the risk estimate is incorrect, an apparent anomaly may simply reflect the correct risk.

In a strong-form efficient market no group of investors should be able to generate excess risk-adjusted returns. Technical analysis, fundamental analysis, and even inside information will provide little value once the information is known.

Tests of the strong form efficient market hypothesis have generally examined the performance of four groups of investors.

1. Corporate insiders
2. Stock exchange specialists
3. Security analysts
4. Professional money managers

Studies of insider buying and selling have provided mixed support for the EMH. At one time, insiders and public investors following insider trades experienced excess risk adjusted returns. However, more recent studies have indicated that public traders can no longer profit after adjusting for transaction costs.

Stock exchange specialists have monopolistic access to certain market data such as unfilled limit orders. Data suggests that specialists are able to earn excess risk-adjusted returns due to their access to this data.

There is some evidence that certain analysts may possess superior information, and that following the recommendations of these analysts may permit excess returns. Often these anomalies appear to be incorporated, which would support the EMH. For example, the Value Line timeliness rating was considered enigmatic as it appeared to consistently predict returns. However, changes in rating are now incorporated in stock prices within a day or two, and transaction costs may limit any usefulness of the anomaly.

In general, tests of professional investors have supported the EMH. On average, such investors do not enjoy superior risk-adjusted returns.

The strong-form efficient market hypothesis assumes that stock prices reflect all information, whether public or private. As such, it encompasses both the weak-form EMH and the semistrong-form EMH. If a market is strong form efficient, it is also weak- and semistrong-form efficient.

In a strong-form efficient market no group of investors should be able to generate excess risk-adjusted returns. Technical analysis, fundamental analysis, and even inside information will provide little value once the information is known.

In essence, the strong form efficient market assumes a perfect market in which all information is cost-free and universally available to all market participants simultaneously.

Secondary markets are those in which securities that have already been issued trade. Transactions occur between investors, and the proceeds do not affect the issuer. Instead, one investor gives another investor cash in exchange for the securities.

Secondary markets provide liquidity to the investors who initially buy the securities. Investors value liquidity because changes in their circumstances may require them to sell the security in order to use the cash for other purposes. By having a liquid secondary market, investors are willing to pay more (accept a lower return) when buying primary issues. This helps issuers raise money at more favorable rates.

Secondary markets also offer issuers price discovery – new issues can be priced according to the value of other similar securities.

Over time, small capitalization stocks have been shown to outperform large-capitalization stocks. However, timing changes in the relative performance between the two groups could lead to still-better performance. In the Fall 2007 Journal of Portfolio Management, L’Her, Mouakhar and Roberge test three nonparametric techniques derived from artificial intelligence and using 20 macroeconomic and financial variables as inputs.

The three approaches are recursive partitioning, a neural network and a genetic algorithm.

Each of the three techniques outperforms a naive small-minus-big strategy, but the best results are derived from taking the consensus of the three techniques.