Generalized beta distributions for describing and analysing intraday stock market data: testing the U-shape pattern

There is an impressive body of empirical evidence which indicates the existence of an intraday U-shaped curve in stock prices. In an effort to shed additional light on the U-shaped curve a new procedure for U-shape testing is introduced. From careful analysis of intraday data it is observed that minimum or maximum stock prices can occur several times during the day. Here, attention is focused on the first time during the day that the maximum or minimum stock price occurred. Because of the importance of the first time during the day that the maximum or minimum stock price occurred, an attempt is made to model these two characteristics with probability distributions. The objective of this study is to use a generalized beta distribution to examine the intradaily behaviour of stocks, using closing stock prices for each one-minute interval, using data from Athens Stock Exchange (ASE). This generalized beta distribution has not been used before to model U-shaped behaviour. The results are consistent with the intraday U-shaped curves, i.e. the time to first maximum (or minimum) stock prices follows a U-shaped pattern. In addition, some potential applications of the generalized beta distribution are discussed and exemplified by analysing the relationship between herd behaviour and U-shaped.     

The inflation-productivity trade-off revisited

Our aim in this paper is threefold. First, to test the robustness of the relation between total factor productivity growth and inflation to the specification of the estimating model; second, to test the stability of their relationship in the short run and in the long run, and third, to investigate the direction of causality between these two variables. To accomplish the first objective, we estimate a generalized Box–Cox cost function using data from the two-digit Standard Industrial Classification of manufacturing industries in Greece during the period 1964–1980. The results show that: (a) the acceleration of inflation from 1964–1972 to 1973–1980 reduced total factor productivity growth in a way that was both statistically significant and sizeable, and (b) even when the effect of inflation is separated from the effects of technical change and economies of scale, the choice of functional form is most crucial. With respect to the second objective, somewhat to our surprise, we find that the inflation-productivity trade-off prevails even in the long run. And, finally, regarding the third objective, it emerges that in the great majority of two-digit manufacturing industries the causality runs from inflation to productivity. On these grounds we conclude that for a precise estimation of the relationship under consideration it is imperative to sort out the three effects involved, do so by adopting the most general flexible functional form for the cost function, and run the appropriate stability and causality tests.

Kmenta's approximation from numerical analysis viewpoint

The paper presents the theoretical problem of Kmenta's approximation to CES production function and studies it from the point of view of numerical analysis. Intervals of overestimation, of underestimation and limitations for the approximation error are obtained.

---

Riassunto Questo lavoro presenta il problema teorico dell'approssimazione di Kmenta alla funzione di produzione CES attraverso la ricerca degli insiemi in cui l'approssimazione è per eccesso e di quelli in cui l'approssimazione è per difetto e la valutazione dell'errore di approssimazione dal punto di vista dell'analisi numerica.

---

---

Versione definitiva pervenuta il 15-3-1982     

Public information as a deterrent to environmental infractions

This paper examines the effect of inflation on total factor productivity growth (TFP), using time-series data for every two digit Greek manufacturing industries. In order to do the above a translog flexible cost function is estimated and used to decompose TFP growth into scale economies, inflation and technical change. The advantage of estimating a very general and flexible cost function is that it allows us, for the first time, to examine empirically a large number of significant relationships between TFP growth and economies of scale, inflation and technical change. The main conclusion drawn from this analysis is that inflation reduces TFP growth in a way, which is sizeable. Furthermore, using standard causality test the direction of causality between inflation and TFP growth was tested at the manufacturing level.     

Testing the efficiency of the Athens Stock Exchange: Some results from the banking sector

In this study we test the efficient market hypothesis in the Athens Stock Exchange for a number of selected stocks from the banking sector. We distinguish between a “weak” and “semistrong” version of the hypothesis depending on the agents' information sets. For the “weak” version we apply a recently developed test by Brock, Dechert and Scheinkman (1987) to test for the presence of nonlinear structure in the residuals of rates of return regressions of these stocks. To test the “semistrong” form of the efficiency hypothesis we carry out tests of cointegration following the methodology of Granger and Engle (1987). We find no noticeable presence of nonlinearties in the standardized residuals for these series. Also we find no evidence of cointegration and hence no Granger causality between the different stocks. Our findings support the “weak” and “semi-strong” versions of the efficient market hypothesis.