Modeling Latent Variance in Production Line Performance Using High-Frequency Data an Application to the Iraqi State Company for Cement Industry and Its Effect on Industrial Output Stability

Abstract

The current research proposes a latent variance in the performance of the production lines as the high-frequency data on operations of the Iraqi State Company of Cement Industry and assesses the impact of this factor on the stability of industrial output. Signals that are constructed on the short-interval production signals are used to construct line performance changes and modeled using GARCH(1,1) to estimate volatility clustering and persistence. The conditional variance is estimated and the result is called the Latent Variance of Line Performance (LVP) and is interpreted as a proxy of operational risk that captures the fact that there is unstable production process. Industrial Output Stability (IOS) is an index of limited stability based on dispersion in achieved output across time, and Unplanned Downtime Ratio (UDR) and Specific Energy Consumption (SEC) are added to moderate reliability shocks and energy intensity. Time-series inference is supported by unit root testing that establishes that the study variables are stationary. The results of quantile regression at τ = 0.30 indicate that LVP has significant negative impact on IOS which suggests an increase in the latent performance volatility decays stability even when the factor of downtime and energy consumption is taken into account. UDR has a huge destabilizing influence, and SEC also has a negative impact, implying that energy inefficiency is accompanied by the unstable regimes of operation. The results indicate that volatility-based indicators can be excellent monitors to signal instability episodes and preventive maintenance and energy management policies to stabilize the output of industries in the cement manufacturing industry.