The overall purpose of this thesis is to extend and apply the Markov Switching Multifractal (MSM) model to various economic problems. To this extent, Chapter 1 lays the ground work for the next chapters by reviewing the MSM model, discussing its properties and outlining its estimation procedures. The chapter also reviews the distributional properties of several commodity markets that make them amenable to the MSM model. Chapter 2 extends the MSM model by incorporating a vector error correction component, which includes in the conditional mean equation, the cointegrating relationship between spot and futures prices. The VECM-MSM model has two distinctive features that incorporate the empirical properties of asset prices. First, it includes an error correction mechanism in the mean equation that incorporates the long-run relationship between spot and futures prices. Second, the model specifies the conditional second moments as a bivariate Markov Switching Multifractal (MSM) model. The VECM-MSM model is applied to study the problem of risk hedging in the futures market. The hedging effectiveness of the proposed VECM-MSM model is evaluated, using a value-at-risk (VaR) approach. Specifically, we compare the hedging effectiveness of the proposed model to those of alternative models by assessing their unconditional and conditional VaR coverages. Models are then ranked in terms of the adequacy and accuracy of their hedged portfolio VaR. The in-sample and out-of-sample hedge effectiveness shows that the VECM-MSM hedged portfolio outperforms alternative hedging strategies in terms of having the lowest rate of VaR violations among the different strategies. Statistical tests of unconditional and conditional coverages also show that the VECM-MSM model better predicts an investor's downside risk in that the VaR predictions are more accurate than the predictions from the alternative models. Chapter 3 of this thesis investigates the excess commodity comovement phenomenon, using the MSM model. One of the stylized facts of commodity prices is their tendency for comovement. The phenomenon implies that seemingly unrelated commodities tend to move together beyond what can be attributed to fundamentals, such as demand and supply conditions, exchange rates, interest rates, industrial production etc. Excess commodity comovement bears significant welfare and risk management implications. For an instance, a synchronous rise in prices of commodities exerts significant inflationary pressure on commodity import dependent countries, and limits their ability to maintain economic stability and resist inflationary pressures. Moreover, to the extent that comovement measures, such as correlation and covariance among commodities, comprise an essential ingredient in risk assessment, pricing, portfolio management and hedging, failure to account for such excess comovement can lead to sub-optimal economic decisions. Therefore within the debate on excess commodity comovement, the objective of this chapter is twofold. First, it analyzes the degree of excess commodity comovement across a variety of commodities. Second, it analyzes the frequency-dependent nature of comovement across related (e.g. crude and heating oil) and unrelated commodities (e.g. copper and corn). First, we find that there is significant comovement between commodity prices, beyond what can simply be explained by macroeconomic fundamentals. Second, decomposing comovements into multiple frequencies, we find that all commodities exhibit long-run excess comovements which are driven by low frequency fundamentals such as weather, demographic and macroeconomic factors. But some commodities also exhibit significant short-run excess comovements that may be attributable to short-run factors such as liquidity constraints, indexation, etc. Third, the dynamic correlations show that excess comovements are higher in periods of high volatility and vice-versa. The final chapter applies a new class of model, the Autoregressive Markov switching multifractal model, for forecasting spot electricity prices. Three variants of the model are examinedEmploying hourly prices from the AESO market, the parameters of the ARX-MSM models are estimated, and one-step-ahead hourly forecasts are obtained. To put the performance of the ARX-MSM models into perspective, the results are compared to those of other notable models used in the literature, namely the AR(1), ARX, ARX-GARCH, mean-reverting jump and the 2-state independent Markov regime switching models. Goodness-of-fit tests indicate that the ARX-MSM models fit the data significantly better than the competing models. Likewise, out-of-sample results show that the ARX-MSM models provide better forecast accuracy.