This paper evaluates whether MIRR is an appropriate criterion for investment decision and the true annual rate of return on capital. Unlike other published papers, the present analysis introduces three important improvements viz. the investment returns are consistent with NCF (NCF-consistent); the two components of returns on capital investment, i.e. return of capital invested (ROC) and return on invested capital (ROIC), are clearly defined and accounted for; and finally, capital amortization schedule (CAS) is used to verify whether the returns are achievable from the NCF generated and therefore NCF-consistent. The appropriateness of MIRR is evaluated using numerical analysis and the main findings are: a. The estimation of MIRR, manually or in excel, is based on the modified net cash flow (MNCF). The MNCF, derived by mathematically adjusting the actual net cash flow (NCF), distorts the intrinsic value of the cash inflow and its timing. With MNCF, the MIRR is lower than the IRR because MIRR failed to fully utilize the NCF generated as shown by the CAS. MNCF is neither NCF-consistent nor accounting concept-consistent (cash vs accrual concept). b. The problem of reinvestment of intermediate income is a fallacy and therefore the MNCF is a meaningless exercise. For the same NCF, the net benefit stream, MIRR is increasing without any limit with increasing investment rates (IR). The NCF is not adequate to support such an increase in MIRR, as revealed by the CAS. Similarly, when the actual IRR is lower than the IR used, the estimated MIRR is higher than the IRR. Again, the NCF is not sufficient to achieve that higher MIRR than the IRR as confirmed by the CAS. This is one of the serious problems with MIRR that is based on the MNCF, a mere data mining exercise. MIRR is not an accurate estimate but a spurious one. c. Again, the problem of multiple IRR is a data problem associated with non-normal NCF. MIRR does not solve this problem either. With non-normal NCF, the cumulative sum of undiscounted NCF is zero or negative or negligible. In those cases, the NCF data leads to multiple IRR. The non-normal MNCF leads to spurious MIRR estimate (also GIRR and AIRR) that is not supported by the actual NCF as revealed by the CAS. Any rate of return must be NCF-consistent. d. With normal NCF also, the MIRR is spurious because of the false reinvestment assumption and the use of MNCF data. The estimated MIRR, based on assumed reinvestment rate, leads to serious problems as explained above. MIRR (when MIRR < IRR) estimate does not fully utilize the benefit stream and leave a closing balance, as revealed by the CAS. Contrarily, IRR fully utilizes the NCF and therefore IRR is higher than the MIRR (paid-off the ROC and ROIC = IRR). When MIRR is higher than the IRR, the NCF does not support that level of MIRR. e. The results of CAS reveal that MIRR is neither the true return nor the annual rate of return. IRR is also not the annual rate of return but it is the true rate of return on the capital remains invested. Both the MNCF and MIRR are not NCF-consistent but may be mathematically-consistent. When there is no intermediate income, the question of reinvestment does not arise. With that type of NCF, the MNCF and the MIRR are NCF-consistent. f. Generalized IRR (GIRR) and Average IRR (AIRR) criteria are also reviewed. They are not NCF-consistent but mathematically generated returns and are based on wrong assumptions (reinvestment). Based on these results, it is evident that the MIRR is a spurious criterion. Investment analysts and decisions makers may wish to move away from using or reporting MIRR as a criterion so also the authors of all published works and finance and economic texts.
Read full abstract