Nowcasting predicts the current situation, the very near future, and even the most recent past. Such models exploit information published earlier than the target variable. In macroeconometrics, nowcasting models are useful tools in the timely assessment of key macroeconomic variables due to publication lags of official statistics in most countries. These models are also used in structural and policy analysis in capturing short-run relationships in the system. In this paper, we formulate a vector autoregressive (VAR) model with mixed frequency data in nowcasting the Philippine economy. We also present a Bayesian estimation approach in addressing over-parametrization on most vector autoregressive models to include larger amounts of variables in the system through different prior specifications. These prior specifications are implemented to cater to violations of certain model assumptions in the VAR modeling framework. We deal with the ragged-edge problem on the data caused by publication lags of official statistics by aggregating the monthly indicators and taking the most recent observation for each series to fill-out gaps from the structure of the database. Forecast evaluation exercise shows better performance of the model in terms of the mean squared forecasting errors and mean absolute errors over benchmark models in the shorter horizon.