A Sequential-trade Microstructure Model with Heterogeneous Information Sets
Enrique Kawamura


I present two cases of sequential trade models of market microstructure with heterogeneous information partitions. First, the trader with information can infer the true value only under certain realizations. The main result is that the equilibrium outcomes look quite different from the literature, suggesting that the information assumptions may be a delicate issue to consider in these models. In the second example I adapt a non-partitional information approach from Geanakoplos (1989). When the informed investor does not observe this signal, she does not infer whether the value of the asset is low or high. Here positive spreads may not lead to trade.

  

Los maleficios que pesan sobre el IVA provincial: Efecto magnificación y paseo de boletas
Osvaldo H. Schenone

Two difficulties arise when different jurisdictions have value added taxes at different rates. First, "magnification effect" which consists of an automatic enlargement of the rate differential over and above the one established by the respective legislations. Second, a fraud known as "invoice sightseeing" which takes place when sales in one jurisdiction are declared as sales to a fictitious customer in a lower rate jurisdiction.

This paper shows that either both or one of these difficulties will unavoidably appear if the jurisdictions adopt different VAT rates. Under certain arrangements, like the zero rate treatment for interjurisdictional transactions, the magnification effect disappears, but the incentives to perform "invoice sightseeing" are still present. Other arrangements are suggested under which such incentives disappear, but the magnification effect does not. Thus, the policy choice is to select a VAT procedure such that its effects upon economic efficiency and revenue are least damaging.

  

Riesgo de Tasas de Interés en Opciones de Intercambio
Salvador Zurita

Margrabe (1978) studied the problem of the valuation of the option of exchanging one risky asset for another. One interesting feature of the valuation formula that he obtained was that the value of the option was independent of the risk-free rate of interest. In this article we extend Margrabe's result to include assets that pay a dividend yield, and commodities (both commercial and precious). In all cases the initial result holds: the exchange option is independent of the risk-free rate of interest.

  

¿Qué Debemos Pedirle a un Candidato? Un Aporte a la Discusión en Torno a la Tasa de Desempleo
Rómulo A. Chumacero


This paper presents three exercises in order to evaluate the discrepancies between the unemployment rate estimated by the Department of Economics of the University of Chile and the National Bureau of Statistics (INE). It is shown that the discrepancies cannot be explained by differences on the questions for eliciting information on employment situation. Furthermore, we show that the historical differences among estimators are significant, but more importantly, that the dynamic relation between both rates and the rate of growth of the economy is also very different. Based on this observation and a theoretical model, we show that the unemployment rate reported by the University of Chile is a more useful candidate for characterizing the referred phenomenon.

  

Day-of-the-week and Size Effects in Emerging Markets: Evidence From Chile
Pablo Marshall and Eduardo Walker


This paper studies empirical regularities of daily log returns for the years 1989 through 1996, using aggregate indexes and quintiles rated by size, for a specific emerging market: the case of Chile.
Within the context of the existing literature on emerging markets, this study's contributions are the following: First, earlier studies use aggregate indexes. This one extends the samples and also considers more detailed information, which gives a better representation of individual stock behavior. Second, non-parametric statistical tests are used as a complements of classical ones.

The study's main result shows important day-of-the-week effects on average returns and traded volumes, but not on variances. These results, obtained with both classical and non-parametric methods, are valid for aggregate indexes, quintiles and sub-periods. We also find a seasonal pattern in the size-effect, which it is significantly positive on Fridays and significantly negative on Mondays. In the case of this emerging market, the evidence is inconsistent with the hypothesis that the weekend effect is due to small-investor-portfolio-adjustment-on-Mondays. Unless there is a reason to believe that bad news is put off to the weekend (and good news to Fridays) especially in the case of smaller firms, the seasonal size-effect and the absence of effects in variances also contradict this hypothesis. There is stronger evidence that favors the hypothesis that investors comply with weekly investment plans, as proposed herein.

Other results confirm that daily returns in the Chilean stock market behave very much like the more developed countries', although the different effects (size-, kurtosis and autocorrelation) are more pronounced. This is also true for the size-based quintiles. Results are also consistent with those obtained by other authors that analyze emerging market monthly index returns.

  

Do investment regulations compromise pension fund performance? Evidence fron Latin America:
A Comment
Salvador Valdés-Prieto


Most of the policy assertions in the paper by Srinivas and Yermo (1999), published in Revista de Análisis Económico, are not backed by the empirical evidence. Their conclusions that "pension funds (in Chile) did not choose an efficient asset allocation or risk return combination", and that "current regulations are severely jeopardizing performance", are not backed by the evidence for Argentina, Chile and Peru. However, these assertions may be valid for the Mexican AFORE, who are banned from investing in equities and foreign securities. We dispute on theoretical grounds their proposal that "investment regulations should not be continued in the long run", by providing benevolent rationales to keep the main investment limits on a permanent basis.

  

Nonparametric estimation of mean and variance and pricing of securities
Akhtar R. Siddique


This paper develops a filtering-based framework of non-parametric estimation of parameters of a diffusion process from the conditional moments of discrete observations of the process. This method is implemented for interest rate data in the Eurodollar and long term bond markets. The resulting estimates are then used to form non-parametric univariate and bivariate interest rate models and compute prices for the short term Eurodollar interest rate futures options and long term discount bonds. The bivariate model produces prices substantially closer to the market prices.

  

Domestic currency emerging market bonds pricing and risk management aspects
Salih N. Neftci

Domestic currency emerging market bonds form an indirect way of trading currency and credit risk. It is true that unlike eurobonds or Bradys, domestic currency emerging market bonds have no default risk in a classical sense. These bonds are issued by local governments and can be paid one way or another by newly issued currency, if need be. Instead, they carry a significant "devaluation risk". In this paper we discuss the environment that supports such bonds, introduce the main parameters and risks, and compare some ways of modelling and pricing these instruments. The paper illustrates that a non-parametric interest rate modelling will be more appropriate for pricing local currency emerging markets bonds.

  

Estimation of a stochastic-volatility jump-diffusion model
Roger Craine, Lars A. Lochstoer, and Knut Syrtveit

 
This paper makes two contributions: (1) it presents estimates of a continuous-time stochastic-volatility jump-diffusion process (SVJD) using a simulation-based estimator, and (2) it shows that misspecified models that allow for jumps, but not stochastic volatility, can give very bad estimates of the true process.

Simulation-based estimation is a very flexible and powerful technique. It is ideally suited to high frequency financial data. It can estimate models with intractable likelihood functions, and since the simulations can be performed in (essentially) continuous-time the estimates are consistent estimates of the parameters of the continuous-time process.

  

Fixed costs and asset market participation
Amir Yaron y Harold H. Zhang


This paper investigates the effects of fixed costs on investor's decision of asset market participation. The model features a continuum of agents with heterogeneous initial wealth and attitude toward risk. We show that under certain conditions there exists a unique competitive equilibrium in which investors optimally choose to stay in autarky, participate just in the riskless asset market or in both the riskless and the risky asset markets. The model is calibrated based on earnings profile from the U.S. We find that using fixed costs that are comparable to the current commission charged by brokers the model can generate participation patterns similar to observed ones. Further, we find participation rates to be very sensitive to the costs differentials associated with entering the risky asset market while relatively less sensitive to the overall levels of fixed costs. Finally, we find that costs make it even harder for dynamic models to replicate the risk free rate and in that sense deepen that puzzle.

  

Quasi-Monte Carlo Algorithm for Pricing Options
Jenny X. Li


The purpose of this paper is to compare the use of Quasi-Monte Carlo methods, especially the use of recent developed (t; m; s)-nets, versus classical Monte Carlo method for valuing _nancial derivatives. Some research has indicate that under certain condition Quasi-Monte Carlo is superior than the traditional Monte Carlo in terms of rate of convergence and accuracy. In particular, theoretic results hinted that the so-called (t; m; s)-net suppose to be the most powerful one among all the Quasi-Monte Carlo methods when the problem is "smooth". However, the application of (t; m; s)-net was not included in the exist-ing simulation literatures. In this paper I will introduce the algorithms of generate the most common Quasi-Monte Carlo sequences, then im- plement these sequences in several path-dependent options. Our in- vestigation showed that Quasi-Monte Carlo methods outperform the traditional Monte Carlo.