Past Price Changes Research Articles

Information Structure, Trading Behavior and Stock Prices in the Thai Market

This study examines the information and trading behavior of investors in the Thai market. This market is an important emerging market in the Pacific Rim, whose structure is different from that of a more developed market. We propose a vector autoregression model to describe and test action and reaction of the portfolio reallocation of investors and the movement of stock prices over time. Using daily market data from January 3, 1995 to October 27 1997 , this study finds that, in the Thai market, the foreign investors bought stocks when prices had risen. This strategy was consistent with a positive autocorrelation in the stock return. The local individual investors bought stocks when prices had fallen, while the local institutional investors disregarded past price changes. These two investor groups also exhibited herd behavior of both informational cascades and interpersonal communications types. They followed each other and reacted negatively to an innovation in the stock return. It is interesting to find that the foreign investors brought new information into the market. The local individual and local institutional investors brought in noise, but the explanatory share of this noisy information in the stock volatility was small. So, the study concludes that the volatility in the Thai market was not excessive.

Endogenous market statistics and security pricing:: An empirical investigation

This study examines empirically the degree to which the history of daytime and overnight price changes and order flow affects estimates of traders' beliefs about future security price changes. Estimates indicate that forecasts of the permanent component of price changes occurring after the open of trading are significantly related to past price changes and order flow; but the same is not generally true for price changes occurring after the close. These results are consistent with models of technical analysis, and models in which the process of trading facilitates price discovery. The evidence also suggests that private information is an important determinant of price movements.

An ordered probit analysis of transaction stock prices

We estimate the conditional distribution of trade-to-trade price changes using ordered probit, a statistical model for discrete random variables. This approach recognizes that transaction price changes occur in discrete increments, typically eighths of a dollar, and occur at irregularly-spaced time intervals. Unlike existing models of discrete transactions prices, ordered probit can quantify the effects of other economic variables like volume, past price changes, and the time between trades on price changes. Using 1988 transactions data for over 100 randomly chosen U.S. stocks, we estimate the ordered probit model via maximum likelihood and use the parameter estimates to measure several transaction-related quantities, such as the price impact of trades of a given size, the tendency towards price reversals from one transaction to the next, and the empirical significance of price discreteness.

Stock Prices and Volume

The authors undertake a comprehensive investigation of price and volume co-movement using daily New York Stock Exchange data from 1928 to 1987. They adjust the data to take into account well-known calendar effects and long-run trends. To describe the process, they use a seminonparametric estimate of the joint density of current price change and volume conditional on past price changes and volume. Four empirical regularities are found: (1) positive correlation between conditional volatility and volume; (2) large price movements are followed by high volume; (3) conditioning on lagged volume substantially attenuates the leverage effect, and (4) after conditioning on lagged volume, there is a positive risk-return relation. Article published by Oxford University Press on behalf of the Society for Financial Studies in its journal, The Review of Financial Studies.

CONSUMER PRICE PERCEPTIONS AND EXPECTATIONS *

This paper examines qualitative quarterly survey data on U.K. consumers price anticipations in the context of a framework designed to allow the authors to test the predictive ability of consumers and the information content of this survey data. The results indicate that the survey data contains useful information but that it is biased in particular ways. Consumers appear to have short memories of past inflation but given this, they appear to use additional information to that contained in recent past price changes in order to forecast future inflation. In particular they may take current short term interest rates into account. Only those households expecting high inflation rates are important in predicting inflation. Copyright 1992 by Royal Economic Society.

Remembering and dating past prices

Two studies investigated memory for past prices. In study 1, 50 West German residents estimated the previous year's prices for both general and specific items rather accurately, but overestimated those of fifteen years previously. These results indicate that misperception of past prices depends on the retention interval rather than the presence of a high rate of inflation. Study 2 asked 65 NZ respondents to date the occurrence of past price changes. Generally, recent price changes were dated too recently, and remote ones too remotely, a result similar to previous price results but dissimilar to those obtained from studies of dating past events, and indicating that biases in past price recall are not the outcome of centring bias.

Past price changes and current trading volume

18 .^ B 0th price and trading volume of common stock are believed to respond promptly to new information. There are, however, no comparable theoretical arguments for a relationship between past information and current trading volume. It is possible, for example, that investors responding to tax incentives may seek to realize capital losses or to avoid realizing gains months after release of information that generated the gains or losses. Thus evidence that this month's trading volume is systematically related to events in previous months can provide insights into further motives for trading. This study uses price patterns in previous months to explain trading volume in the current month. We find that past price changes do influence current incentives to trade for both tax and non-tax motives. Volume tends to be higher than normal for winners (stocks that performed well in the prior period) and lower than normal for losers. The influence of tax-related motives is shown by seasonal variations in the relationship between past price changes and volume. Specifically, the trading volume of losers is higher than normal in December, and winners show a relatively higher trading volume in January. Our methodology does not permit us to distinguish clearly among the various non-tax motives for trading. Some of these motives are consistent with customary ideas about rational behavior; others are not. This study is based on Lakonishok and Smidt [198#6], which can be consulted for technical details [

Wages Policy and Unemployment in Australia

It is generally accepted among Australian economists that high rates of growth of wages at various times over the last decade have made the achievement of full employment and lower rates of inflation more difficult. Fear of a further acceleration of wage growth has also been a policy constraint. It is when one attempts to go further that considerable disagreement becomes evident. From a labour market viewpoint, there does not seem to be a consensus as to the answers to the two broad questions that should be addressed. First, how important are wage changes relative to other factors that depress employment and add to the inflation rate? Second, how does government policy impinge on the process that generates wage changes? To do justice to both of these in a paper as short as this one is difficult, so I shall focus most of the paper on the second question. The relationships between government policy and wage changes are particularly important at this time. For most of the last decade, successive governments have operated a national wages policy, which to a large degree has been based on the indexation of wages for past price changes. The latest experiment, the Prices and Incomes Accord, is the centrepiece of the current government's employment and inflation policy, and, after almost two years of wide community and media support, the key principles upon which it is based are suddenly being questioned.1 Our prime endeavour, therefore, will be to build up sufficient understanding of the wage determination process so that some evaluation can be made as to the effectiveness of these national wage policies. This turns out to be a remarkably difficult task which generates considerable concern in our mind that the data may not be sufficiently rich to serve as a basis for sound judgments. It does appear, however, that the effect of a given level of unemployment upon wage changes has consistently weakened over the last decade. Why this should be so is a major unresolved puzzle. I-offer a loosely constructed theory that is consistent with the facts, but it is by no means the only possible explanation. Section I provides a description of wage-setting in Australia and some key facts of the economy. In Section II the statistical analysis of wage changes focuses on two competing theories: one is the Phillips curve approach, and the other stresses that wage outcomes may be, to a significant degree, independent of the level of unemployment and may depend instead on utilization rates of labour -within the firm. Section III considers the efficacy of incomes policies in Australia. Some final comments areoffered in Section IV.

Price Inertia and Policy Ineffectiveness in the United States, 1890-1980

This paper introduces a new approach to the empirical testing of the Lucas- Sargent-Wallace (LSW) ineffectiveness proposition. Instead of testing that hypothesis in isolation from any plausible alternative, the paper develops a single empirical equation explaining price change that includes as special cases both the LSW proposition and an alternative hypothesis. The alternative, dubbed NRH-GAP, states that prices respond fully in the long run, but only gradually in the short run, to nominal aggregate demand disturbances. A second innovation is the development of a quarterly data file for the period 1890-1980, thus opening up more than 200 new quarterly observations for analysis. A third innovation is the testing of three different methods of introducing persistence effects into the LSW analytical framework. In conflict with the predictions of the LSW approach, the results here exhibit uniformly high coefficients of real output and low coefficients of price changes in response to anticipated nominal GNP changes. Further, price changes respond positively and output responds negatively to lagged changes in prices, reflecting the short-run inertia in price-setting that forms the basis for the alter- native NRH-GAP approach. Evidence is also provided that velocity tends to respond negatively to anticipated changes in money, in contrast to the usual assumption in this literature of random serially independent velocity changes. Two shifts in the structure of the price-setting process are noted--a much higher degree of price responsiveness during World War I and its aftermath, and a longer mean lag in the influence of past price changes after 1953. Of independent interest, beyond its treatment of the policy ineffectiveness debate, is the treatment in the paper of changes in monetary regimes, and of the impact of programs of government intervention. The money creation process exhibits a highly significant change in structure before and after World War I, and a marginally significant change in 1967. The results identify five episodes of government intervention that significantly displaced the time path of prices -- the National Recovery Act of 1933-35, and price controls during the two world wars, Korea, and the Nixon era.

The Term Structure of Inflationary Expectations and Market Efficiency

FISHER'S [15] ORIGINAL FORMULATION OF an equilibrium relationship in which a nominal interest rate is equal to a real rate of interest plus the expected rate of inflation has stimulated considerable interest and research. Although many empirical studies have utilized a single equation distributed lag specification1 of past price changes (which might be induced from a rational expectations argument) as a proxy for expectations, T. F. Cargill and R. A. Meyer [5] have emphasized the necessity of a simultaneous equation structure if such distributed lags were to be employed. An alternative approach using Livingston's direct observations on expectations and allowing for the possibility of incomplete adjustment to expectations was employed by Cargill [2], W. Gibson [16] and D. Pyle [22]. These evolving improvements in conception of the empirical problem, data and econometric approach have led to a particularly interesting paper by K. Lahiri [18] in which S. Turnovsky's [25] expectations formation equations are coupled to the original Fisherian equation and then appropriate simultaneous equation techniques are applied to the resulting econometric relationships. From a slightly different perspective, E. Fama [13] has emphasized the implication for market efficiency embedded in the interest rate and expected inflation relationship. All analyses to date rest on the assumption that the real rate of interest is constant over time. Thus, as Fama noted, empirical implications for market efficiency tests really constitute a test of the joint hypothesis of efficiency and a constant rate.2 Despite the advances in conception of the empirical problem, data improvements, and improvements in econometric technique, the question of variation in the real rate of interest and the possibility of incomplete adjustment to changes in the real rate as well as to expected inflation leaves the extent and stability of the Fisherian relationship far from a settled matter. In addition, a much richer class of hypotheses emerges when Fama's emphasis on the market equilibrium aspects is carried further. No prior work has utilized any type of data to estimate the terni structure of expectations which is the key to Fisher's hypothesis when applied across maturities. In addition, the contributions of W. Sharpe [24] and J. Lintner [20] in providing an equilibrium capital asset pricing theory can be

Comparing Survey and Rational Measures of Expected Inflation: Forecast Performance and Interest Rate Effects

THE MEASUREMENT OF EXPECTED INFLATION continues to be an important and perplexing issue in empirical studies of macroeconomic relations. Although most researchers have followed Fisher [10] in assuming that a distributed lag of past price changes serves as an adequate proxy for anticipated inflation, several recent investigations employ instead direct observations on inflation forecasts from surveys. Simultaneously with this work, there has been a growing emphasis, particularly in the theoretical literature, that expectation formation should conform to the notion of rationality advanced by Muth [17]. A potential conflict thus arises if the survey data violate the rationality hypothesis. The main survey source is the Livingston survey of professional economists, which iIlcludes their forecasts of future price levels. These data have often been used in empirical tests of the effects of anticipated inflation on interest rates [ 1 1, 19, 13, 5]. Tests by Pesando [18] and Carlson [6] the latter using his reworking of the survey responses indicate, however, that the survey data are indeed in-

INTERTEMPORAL STABILITY OF THE RELATIONSHIP BETWEEN INTEREST RATES AND PRICE CHANGES

WHILE THE FIRST RIGOROUS STATEMENT of the relationship between interest rates and anticipated price charges is due to I. Fisher [9] over seventy years ago, the subject has been most enthusiastically studied during the last decade. A partial listing of recent studies would include: L. C. Anderson and K. M. Carlson [1], T. F. Cargill and R. A. Meyer [6], Cargill [4], M. Feldstein and 0. Eckstein [8], W. E. Gibson [10] [1 1] [12], M. J. Hamburger and W. L. Silber [13], D. Pyle [17], T. J. Sargent [18], and W. P. Yohe and S. S. Karnosky [21]. Fisher's original formulation was expressed in terms of anticipated price changes; however, the lack of a generally accepted measure of anticipated prices has lead to the widespread use of distributed lag functions of past price changes as a proxy for anticipated price changes. Thus the majority of studies on the Fisherian relationship really embed two hypotheses: (1) that actual past price changes embody information about anticipated price changes and (2) anticipated price changes influence nominal rates of interest. With the exception of the papers by Gibson, Pyle, and Cargill, studies of the Fisherian relationship have only provided indirect rather than direct evidence. The majority of studies have lead to several widely accepted implications about the Fisherian relationship. First, there is a significant statistical relationship between past price changes and interest rates. Adjustment coefficients are often less than one and tend to vary inversely with the maturity of the interest rate. Second, the positive relationship between interest rates and price changes is characterized by a long distributed lag response of interest rates to price changes. The only exception to this finding occurs for the period from 1950 through 1970 where lags of less than two years have been observed. Third, despite the long periods of time studied, variety of data bases, and estimation procedures employed, many studies concluded that past price changes account for a significant amount of the variation in rates of interest as measured by the computed R2 and that the regression coefficients are both positive and significant at conventional significance levels. The state of the empirical evidence is best summed up by Gibson in a recent paper: .....while the matter was in doubt just a few years ago, it now appears true that interest rates have indeed responded to changes in price expectations in the United States. [12, p. 753]. The extensive number of studies on the Fisherian relationship may be open to

Earnings Growth and Price Change in the Same Time Period

FREQUENTLY the most important determinant of the rate of return in equity investments is the rate and direction of price change.' Success in investments, if it not be due to chance, is largely a function of the ability to predict price changes. The importance of predicting price changes has led to a number of studies aimed at discovering the determinants of price change. These studies may be conveniently classified into three groups: first, those studies that sought to predict future price changes from past price changes; second, those investigations that attempted to predict price changes from price ratios, such as the price/ earnings ratio; and third, those studies which sought to predict price changes from past changes in other variables, such as earnings. In the last decade a score of studies were made which attempted to discover whether future price changes could be predicted from past price changes. The results were generally disappointing. It was found that successive price changes tended to be independent; the price change of a stock in one period had little bearing on the price change in the next period. Moreover, the relative price change of a stock (relative to other stocks) in one period was not indicative of the relative price change of that stock in the next period.2 Although the full import of the results of these studies is not yet clear, the studies certainly have important implications for financial analysis. Some have concluded that the results call into question the very utility of fundamental financial analysis.3 This conclusion was certainly premature and may be questioned, as will be shown below. The results of the second and third groups of studies were only partly encouraging. The correlation between price ratios and future price changes, though promising, was frequently neither significant nor positive and price changes in one period tended to be independent of earnings changes in the preceding period.4 None of the studies were devoted to the question of the relation between relative earnings changes and relative price changes in the same period. This question should probably have been examined first. Even though previous e arnings growth was unrelated to present changes in prices, perhaps percentage changes in prices and earnings in the same period were highly correlated. Per share earnings growth could still have a substantial influence on simultaneous price changes. If there were little connection between relative earnings changes and p r i c e changes in the same period, then the ability to predict relative earnings changes might be of limited value. If, on the contrary, there were high correlation between relative earnings changes and relative price changes in the same period, then the ability to predict relative earnings changes would be extremely important.5 The purpose of this article is to report the results of a study of the influence of rates of growth of per share earnings on percentage changes in stock prices in the same period.

Interregional Competition in Agriculture. With Special Reference to Dairy Farming in the Lake States and New England.

THE purpose of this study was to forecast, from the mid nineteen-thirties, the probable trends in dairy farming in two districts of the United States which were competitors in certain markets for liquid milk and milk products. Because adjustments in output take time, it was assumed that ten years would be required before farmers in these regions had fully adjusted their plaiis to changes in the relative prices of milk and other products. The method adopted was to select certain farms regarded as of the main systems of dairy farming in the chosen regions. For each of these farms, a series of budgets was drawn up showing the most profitable combination of enterprises and level of output, assuming that milk prices had changed, relative to other prices, by a stated proportion, either up or down. Such calculations involved a detailed study of each farm, its current trends in output which might be reflecting past price changes, its limitations in the way of building, soil, climate or labour, and the range of opportunities open to it at differing price levels. The forecasts made for these typical farms were then adjusted to allow for the average of a large group in which there were bound to be casualties by storm, disease or change of occupier; a supply curve) was then constructed which showed what levels of milk sales could be expected from each farming area and from each region ten years after there had been a substantial change in the relative level of milk prices. Finally, the actual sales from these regions and as far as possible from the originally surveyed farms, were obtained for 1946 to compare with the forecasts made before the war. Over this period, the change in milk prices was not much below that assumed as the basis for one set of calculations-a relative rise of 20%, but the actual level of output turned out to be much higher than had been deduced from the pre-war studies. The discrepancy between actual and forecasted sales differed in size between regions, between counties and between the budgeted farms, but the forecasts were all too low. The original calculations had assumed, naturally, that there would be no appreciable change in technique, but this particular decade brought two