11. statistical analysis Return to the outsourcing illustration in Tables 11.3 and 11.4. Everything remains as specified, and the prices of the products are
P1 = 600 and
P2 = 1, 100. The new feature concerns whether the overhead LLA is well specified. To this end, the following data from the 10 most recent periods are extracted from the accounting library.

t OVA OVS DLA DLS DMA DMS 1 1,428 1,596 256 76 24 612 2 1,811 2,228 446 106 49 985 3 1,775 2,306 428 78 61 986 4 1,005 2,239 205 25 86 1,003 5 1,687 1,701 404 54 113 676 6 1,568 2,502 365 15 130 1,028 7 1,299 2,256 262 82 37 1,066 8 1,625 2,268 366 26 79 1,016 9 1,570 2,405 385 45 122 1,069 10 1,411 1,656 234 98 42 679 Notice we have two distinct overhead pools, one for assembly (OVA)

and one for subassembly (OVS). No other overhead is present, so total overhead is simply OV = OVA +OVS. The other cost pool totals refer, of course, to direct labor and direct material in the assembly and subassembly spheres. No other direct costs are present.

(a) Using these data, regress total overhead on total direct labor cost. Is your resulting LLA estimate statistically consistent with that assumed originally?

(b) Using this newly minted overhead LLA, determine the implied contribution margins and optimal solution when the outsourcing option is available (with price P = 250 of course).

(c) How much would the firm pay to distinguish between the original and your new statistically estimated LLA?

(d) Further exploration suggests overhead in the subassembly area is best related to direct material (DMS), while overhead in the assembly area is best related to direct labor (DLA). Performing the implied regressions, are the data consistent with this conjecture?

(e) Now what happens, using the two overhead LLAs to the contribution margins and optimal solution?

(f) Why does the projected profit vary as you move among the three possible specifications of the overhead structure?