Search and herding effects in peer-to-peer lending: evidence from prosper.com

I examine loan data from Prosper.com—a website which allows borrowers to post loans and for lenders to bid on those loans. The Prosper market somewhat resembles the theoretical model of search, herding, and crowding in a large market described in Berkovich and Tayon (Phd. dissertation—Essays on search and herding. University of Pennsylvania, Pennsylvania, 2009). That model predicts that assets with high and low prices have high variance in the difference between price and true value. These extreme price regions of the asset-space are where private information provides excess returns. In the Prosper market, I find some evidence for the model since loans with low ex post returns show higher variance of the difference between price and ex post return. I also find that high-priced loans provide excess returns even after accounting for risk-aversion.     

Implications of the Discreteness of Observed Stock Prices

Stock prices on the organized exchanges are restricted to be divisible by ?. Therefore, the “true” price usually differs from the observed price. This paper examines the biases resulting from the discreteness of observed stock prices. It is shown that the natural estimators of the variance and all of the higher order moments of the rate of returns are biased. An approximate set of correction factors is derived and a procedure is outlined to show how the correction can be made. The natural estimators of the “beta” and of the variance of the market portfolio, on the other hand, are “nearly” unbiased.     

Imperfect Information and Cross-Autocorrelation among Stock Prices

I develop a model to explain why stock returns are positively cross-autocorrelated. When market makers observe noisy signals about the value of their stocks but cannot instantaneously condition prices on the signals of other stocks, which contain marketwide information, the pricing error of one stock is correlated with the other signals. As market makers adjust prices after observing true values or previous price changes of other stocks, stock returns become positively cross-autocorrelated. If the signal quality differs among stocks, the cross-autocorrelation pattern is asymmetric. I show that both own- and cross-autocorrelations are higher when market movements are larger.     

THE DETERMINANTS AND DYNAMICS OF BID-ASK SPREADS ON THE LONDON STOCK EXCHANGE

We analyze the effect of various factors on the size of spreads on the London Stock Exchange since “Big Bang” and find that the price of a security, volume of transactions, risk associated with security returns, and degree of competition among market makers explain 91 percent of the cross-sectional variation in spreads. The results are consistent with the argument that the inside spread encompasses the order-processing, inventory-adjustment, and adverse-information cost of spreads. We also investigate the speed at which spreads move toward their normal levels after a temporary deviation. Although the speed of adjustment varies across firms, the cross-sectional median of 0.896 indicates it takes more than one period (day) for the adjustment to be completed. The volume of transactions and the degree of competition among market makers are the significant factors that affect the speed of correction in spreads toward their normal levels. This implies private information is incorporated more quickly into prices for stocks with greater competition and high trading volume.     

Flight-to-safety and retail investor behavior

Flight-to-safety (FTS) episodes are associated with substantial yet short-lived changes in expected returns on equities and bonds. These price changes are typically surrounded by active trading and/or risk transfer between different investors. Using aggregate net exchanges, flows from bond to equity mutual funds, in the US for a period 1984 until 2015, I empirically investigate retail investor behavior around FTS episodes. Overall, I find a reversal statistical relation between net exchanges and market excess returns. A one-standard-deviation shock to net exchanges leads to an increase of market excess return of 1.75%, of which 72% is reversed within 5 months. In particular, FTS episodes are preceded by periods where more risk averse investors, e.g. retail investors, rebalance their portfolios towards risky assets. A trading strategy that is based on signals from past net exchanges outperforms the market portfolio, significantly during FTS periods by 1.4% monthly. However, I find that the observed reversal relationship is not necessarily due to price ‘noise’ induced by uninformed trading. It is more reasonable that the sudden increase in market stress and selling of risky assets is caused by other demand/supply shocks driven by increased economic uncertainty.     

Does option trading convey stock price information?

After executing option orders, options market makers turn to the stock market to hedge away the underlying stock exposure. As a result, the stock exposure imbalance in option transactions translates into an imbalance in stock transactions. This paper decomposes the total stock order imbalance into an imbalance induced by option transactions and an imbalance independent of options. The analysis shows that the option-induced imbalance significantly predicts future stock returns in the cross section controlling for the past stock and options returns, but the imbalance independent of options has only a transitory price impact. Further investigation suggests that options order flow contains important information about the underlying stock value.     

The market reaction to cross-listings: Does the destination market matter?

This paper examines (i) whether market reactions to cross-listings differ across destination markets and (ii) to what extent the following explanations for value creation around cross-listings can account for differences in market reactions across cross-listings on various destination markets: overcoming market segmentation, increased market liquidity, improved information disclosure, and better investor protection (“bonding”). We analyze 526 cross-listings from 44 different countries on eight major stock exchanges and document significant announcement returns of 1.3% on average for cross-listings on US exchanges, 1.1% on London Stock Exchange, 0.6% on exchanges in continental Europe, and 0.5% (not significant) on Tokyo Stock Exchange. We find evidence consistent with improved disclosure and bonding creating value for cross-listings on US exchanges, while overcoming segmentation and bonding are associated with higher announcement returns on the London Stock Exchange. The evidence is mixed for continental European exchanges and for Tokyo. Our results highlight the role of the destination market in value creation around cross-listings.     

Price experimentation and security market structure

We examine the role of market makers in facilitating price discovery.We show that a specialist may experiment with prices to inducemore informative order flow. thereby expediting price discovery.Market makers in a multiple-dealer system, unlike a specialistsystem, do not have the incentives to perform such costly experimentsbecause of free-rider problems. Consequently, the specialistsystem may provide open markets where competition fails butat the cost of wider bid-ask spreads. We analyze the effectof experimentation on the bid-ask spread and provide an exploratoryanalysis of intraday specialist data that is consistent withour price experimentation hypothesis.     

The market price of risk of the variance term structure

In this paper I examine the market price of risk of the variance term structure. To this end, the S&P 500 option implied variance term structure is used as a proxy for aggregate variance risk. Principal component analysis shows that time variation in the variance term structure over the 1996–2012 period can be explained mainly by two factors which capture changes in the level and slope. The market price of risk of each factor is estimated in the cross-section of stock returns. The slope of the variance term structure is the most significant factor in the cross-section of stocks returns and carries a negative risk premium. The slope factor has also some predictive ability over long horizon equity returns.     

Long-term dependence in common stock returns

The efficient market, martingale model of security price movements requires that the arrival of new information be promptly arbitraged away. A necessary and sufficient condition for the existence of an arbitraged price is that statistical dependence among prices must decrease very rapidly. If persistent statistical dependence is present, the arbitraged price changes do not follow a martingale and should have an infinite variance. Using a technique for detecting long-term dependence, called R/S analysis, 200 daily stock return series are studied; many series are characterized by long-term dependence. Thus, in the presence of long-term dependence, the martingale model does not hold. Also, the distribution of security returns is non-normal stable Paretian as opposed to Gaussian.     

Related securities and price discovery: Evidence from NYSE-listed Non-U.S. stocks

In this paper we analyze whether handling related securities improves a market maker's information environment and helps to incorporate new information in stock prices. Our empirical tests are focused on New York Stock Exchange specialists and the U.S. share in price discovery of 64 British and French companies cross-listed on the NYSE. We define related securities as stocks from the same country, the same region, or other foreign stocks. We find strong evidence that a higher prominence of related stocks in the specialist portfolio is associated with a higher U.S. share in price discovery of our sample firms. We interpret our findings as evidence that concentrating market makers in similar stocks reduces information asymmetries and improves the information environment as market makers can extract information relevant to a stock from order flow to related securities. To support our argument, we show that the adverse selection component of the bid–ask spread is negatively related to the prominence of other foreign stocks in the specialist portfolio.     

Microstructure and reverse stock splits

In this paper we show that, similar to NYSE/AMEX stocks, NASDAQ stocks exhibit significant ex date returns for reverse stock splits. Although the 10-day cumulative return after the ex date is close to –10%, this does not violate market efficiency, because the average bid-ask spread for the reverse split stock is at least double this return. We also document that these large negative returns are mostly due to a drop in the ask price while bid prices barely change at all. Furthermore, the ex date returns are negatively related to trading volume.These results suggest that there is abnormal selling and a significant buildup of market makers' inventories near the ex date. To reduce the inventory buildup, market makers lower ask prices to induce buying by investors, resulting in the observed negative returns. Lowering bid prices, an alternative strategy for reducing inventories, is not attractive to market makers due to competitive factors and the reduction of commissions associated with a smaller number of transactions. Notably, selling investors have no incentives to sell their stocks early to avoid the observed negative ex date return, since this return is largely an ask price phenomenon and does not represent realized returns to sellers.     

The Risk‐Return Relation in International Stock Markets

We investigate the risk‐return relation in international stock markets using realized variance constructed from MSCI (Morgan Stanley Capital International) daily stock price indices. In contrast with the capital asset pricing model, realized variance by itself provides negligible information about future excess stock market returns; however, we uncover a positive and significant risk‐return tradeoff in many countries after controlling for the (U.S.) consumption‐wealth ratio. U.S. realized variance is also significantly related to future international stock market returns; more importantly, it always subsumes the information content of its local counterparts. Our results indicate that stock market variance is an important determinant of the equity premium.     

THE SIZE EFFECT IS PRIMARILY A PRICE EFFECT

Previous research shows that stock returns are related to firm market value and earnings yield. This study decomposes the measurements of market value and earnings yield into separate components (share price and shares outstanding for market value, earnings per share and share price for earnings yield) and examines the relationship between stock return and these components. Share price represents approximately three-fourths of the relationship between stock returns and either market value or earnings yield. Potential causes for this phenomenon are advanced.     

The Role of Tick Size in Upstairs Trading and Downstairs Trading

This paper examines the impact of reducing the tick size on market-making behavior on The Toronto Stock Exchange. The results indicate a significant decrease in the percentage of trades of fewer than 10,000 shares involving the upstairs traders and a significant increase in the percentage of trades of fewer than 1,000 share involving the designated market makers. Consistent with this finding, the upstairs traders earn significantly lower returns on non-block trades and the designated market markers earn lower returns on trades smaller than 1,000 shares. We conclude the tick size reduction benefits the trading public.Journal of Economic LiteratureClassification Numbers, G20, G24.     

Asset pricing,time-varying risk premia and interest rate risk

This paper investigates the role of interest rate risk in explaining security price changes. We develop and test a two-factor linear beta pricing model of security returns in which the factors are the excess returns on the long-term, riskless bond and the equal-weighted equity market index. We find that time-variation in the interest rate and market risk premia influence expected security returns. Furthermore, conditional interest rate volatility affects security returns, particularly during periods of substantial interest rate movements.     

Decentralized exchanges: The “wild west” of cryptocurrency trading

Cryptocurrencies are traded on two types of exchanges – centralized and decentralized. Although trading in the largest cryptocurrencies primarily occurs on centralized exchanges, most newly issued tokens can only be exchanged using decentralized platforms. Volumes in these decentralized exchanges (including automated market makers) has recently increased exponentially. We examine the role of this new and unmonitored market, utilizing an extensive sample of tokens exclusively traded on decentralized platforms. We show significant differences in the listing and trading characteristics of these tokens relative to their centralized equivalents. A small selection of these tokens obtain listing on a centralized exchange during the sample period, which is accompanied by a significant increase in trading activity, consistent with market segmentation. A centralized listing results in a migration of volume away from decentralized platforms, revealing a strong preference by tokenholders for deeper and more liquid markets over the increased security and anonymity offered by decentralized exchanges.     

Seasonals,price adjustment delays and estimated systematic risk

This article analyzed potential interactions between seasonals and price adjustment delays on estimated systematic risk. It was shown that seasonals in unobservable true security returns can induce inconsistencies into the generalized Scholes and Williams estimator of systematic risk. An alternative estimator was proposed that is consistent in the presence of seasonals in the unobservable true returns. The direction of induced bias is unpredictable a priori, thereby representing a potentially important research consideration in market efficiency tests using abnormal returns. NASDAQ and Dow Jones 30 Industrial return data for the period 1983–87 were used to evaluate the proposed estimator against the OLS and generalized Scholes and Williams (GSW) alternatives. The absolute difference between the GSW and our estimator, that is the seasonal-induced bias, for NASDAQ stocks was negatively correlated with market capitalization. Moreover, seasonal-induced bias was larger for NASDAQ stocks than more highly capitalized Dow stocks. These empirical findings indicate that seasonals and price adjustment delays can interact to bias estimated systematic risk, where price adjustment delays would be projected to be more acute for smaller capitalization stocks.     

Stock prices and the flow of information in the Athens Stock Exchange

The paper investigates the dynamics of price changes and information flow to the market in the Athens Stock Exchange in Greece using daily data over the period 1988 to 1993. A generalised autoregressive conditional heteroskedastic (GARCH) model in stock returns is shown to reflect time dependence in the process generating information flow to the market. Using daily trading volume or value as proxies for information flow, we find them to be significant in explaining the variance of daily returns and to reduce GARCH effects substantially. This has implications for the informational efficiency of the market.     

On the theory of market information efficiency

This paper discusses the concept of market information efficiency and demonstrates the following: As the number of traders who participate in the market becomes large, the variations in the price of a security caused by the variations in traders' beliefs make the market price vary as if traders all knew the ‘true’ distribution of returns on the security. The empirical implications of the analysis are also explored.