Peter S. Sephton

Tests of exchange market efficiency: fragile evidence from cointegration tests

Abstract Several recent papers have employed Engle and Granger (1987) cointegration tests to determine whether or not markets are efficient. We use a new test for cointegration due to Johansen (1988) to examine market efficiency with respect to the US dollar exchange rates for Canadian, Japanese, and West German currencies. The fragile nature of the test indications suggests that cointegration methodologies cannot be relied upon to provide reliable evidence on market efficiency.

Response surface estimates of the KPSS stationarity test

Abstract A series of Monte Carlo experiments provide the basis for response surface estimates of approximate critical 1%, 5%, and 10% values of the Kwiatkowski et al. test of stationarity. The methodological approach closely follows MacKinnon.

Spatial Market Arbitrage and Threshold Cointegration

Goodwin and Piggott reported that corn and soybean prices in spatially separated markets in North Carolina exhibited threshold cointegration and that commodity prices in different markets may persistently diverge. Here, a multivariate approach is used to test for threshold cointegration and nonlinear cointegration. The results suggest that departures from the law of one price do not persist indefinitely.

Optimal hedge ratios at the Winnipeg commodity exchange

Multivariate GARCH models are employed to estimate time-varying hedge ratios for three commodities traded on the Winnip eg Commodity Exchange. GARCH hedge ratios are shown to be superior to those based on the traditional regression approach to calculating th e optimal hedge.

A note on the efficiency of the London metal exchange

Abstract In this note we examine the market efficiency hypothesis with respect to six metals traded on the London Metal Exchange (LME). Using overlapping data and both single and multimarket models, we find evidence contradictory to the tenet that the LME is an efficient market.

The efficiency of the london metal exchange: another look at the evidence

Recent evidence suggests that some markets on the London Metal Exchange do not exhibit the major characteristics of efficient markets: rationality and risk neutrality. In this note we ilustrate several shortcomings of recent work in this area. We suggest that one can be confident that the tin market exhibited a risk premium, that is, was inefficient, between 1976 and 1985. Reliable results on the efficiency of the lead, copper, aluminium, and nickel markets escape the Fama approach to testing the efficient markets hypothesis as it is currently employed.

Hedging wheat and canola at the Winnipeg Commodity Exchange

Data on feed wheat and canola contracts traded on the Winnipeg Commodity Exchange is used to evaluate the benefits to constructing hedge ratios using multivari-ate GARCH models.

Fractional cointegration: Monte Carlo estimates of critical values, with an application

A fractionally integrated series is mean-reverting, and may be covariance stationary. Recent interest in fractional integration has been extended to tests of whether series are fractionally cointegrated. This article provides simulated critical values for use in tests of fractional cointegration for up to six variables, over a number of sample sizes. An example illustrates the potential merits of tests for fractional cointegration.

Cointegration tests on MARS

Multivariate adaptive regression spline (MARS) models due to Friedman (1991) are employed to examine non-linear cointegration. Critical values of the Dickey-Fuller cointegration test statistics, appropriate to the MARS model, are presented. Several empirical examples demonstrate the gains to the non-linear modelling of economic time series.

On interest rate innovations and anticipated monetary policy

Abstract In this note I investigate whether anticipated money matters by following McCallums (1983) suggestion that nominal interest rates be included in a VAR system estimated on data spanned by a policy of interest rate targeting. The evidence is in favour of Gordons (1982) ‘natural rate hypothesis — gradual price adjustent’ model of policy ineffectiveness. Ambler and Phaneuf (1988) have shown the results can be interpreted as evidence in favour of contract-based models of the business cycle.