Exploring Inventory Leanness on Firm Performance with a Non-linear Empirical Leanness Indicator using Kernel Regularized Least Squares (KRLS) Regression

The extant literature assumes linearity between inventories and sales We relax this assumption, suggesting the relationship may indeed be non-linear. We develop a measure of leanness termed non-linear empirical leanness indicator (NELI) using kernel regularized least squares (KRLS), which is an efficient and interpretable machine learning method allowing us to accurately capture the true functional form in the data without losing any explanatory power offered by traditional regression models. Next, we benchmark and demonstrate its potential as a better explanatory variable than the traditional measures in explaining both accounting-based performance measures (ROA) as well as market based measures (Market Capitalization). Prior literature tends to bunch industries together potentially missing industry specific characteristics. To avoid this, we analyze this relationship accounting for firm-specific fixed effects which encapsulates any industry specific effects. We examine this relationship in an emerging economy - Indian manufacturing firms. The dataset we use has been used in publications such as Econometrica, Journal of Finance and Strategic Management Journal amongst others, alluding towards its reliability. Our results suggest that overall leanness has a positive impact on firm performance. We also find that this effect is more pronounced in firms operating in a relatively stable environment. Overall, our results contribute to the literature by: a) documenting a positive impact of leanness on performance; b) showing that the effect can be dependent on the underlying business industry dynamics; and, c) showing that this effect is documented in samples outside the western world where management theories and practices are evolving.