Zero-power reactor noise is useful for subcriticality measurements. Based on the nuclear reactor physics and the theory of neutron detection, this paper theoretically clarifies that the third- and fourth-order neutron correlation factors Y_3 and Y_4 can be expressed as functions of the second-order neutron correlation factor Y. In particular, if the neutron-counting gate width is sufficiently large, the saturation values Y_3⁄Y^2 and Y_4⁄Y^3 are almost equal to the unique combination numbers, ‘3’ and ‘15,’ for a source-driven subcritical system, where the subcriticality is less than 10000 pcm. These unique combination numbers, ‘3’ and ‘15,’ for Y_3⁄Y^2 and Y_4⁄Y^3 were validated using actual zero-power reactor noise measurements carried out at the Kyoto University Criticality Assembly. In this study, the estimation of statistical errors and correlations between different gate widths owing to the bunching method was achieved by the moving block bootstrap method. For a sufficiently long measured reactor noise in a steady and unperturbed state, a statistical test for the evaluation of the critical state and the absolute measurement of subcriticality can be carried out by statistically quantifying the difference between the measurement value of Y_3⁄Y^2 and the unique combination number.