Firm strategy and age dependence: a contingent view of the liabilities of newness, adolescence, and obsolescence

Administrative Science Quarterly,  June, 1999  by Andrew D. Henderson

• E-mail
• Print
• Link

Failure-rate estimation. Failure rates were assessed using both discrete-time and continuous-time event history analyses (Allison, 1984; Tuma and Hannan, 1984). The data from IDC were updated annually, so the year in which a firm failed was known, but not the precise month or day. Discrete-time, maximum likelihood models are designed to handle this (Allison, 1982, 1984). Obviously, a firm might fail at any point during the year, so the objective was to recover what were actually continuous-time hazard rates. Discretetime models that use a complementary log-log function accomplish this by accounting for both (a) the discrete nature of the available data and (b) the continuous nature of actual failure processes (Allison, 1982; Petersen, 1991). These models were estimated using the LOGISTIC procedure in SAS with the LINK = CLOGLOG option.

Table 1

Hypothesized Effects Using Terms in Equation 1(*)

Hypothesis                           Predicted effects

1a. In the years shortly after       ([[Beta].sub.3] [Tau] +
founding, age will produce lower     [[Beta].sub.4] [[Tau].sup.2])
rates of sales growth in             [less than] 0, for 0 [less
proprietary strategists than in      than] [Tau] [less than or equal
standards-based strategists.         to] 4.

1b. For standards-based              [[Beta].sub.1] [greater than]
strategists, rates of sales growth   0; [[Beta].sub.2] [less than]
will have an inverted U-shaped       0; and an inflection within the
relationship with age.               observed range: 1 [less than] -
                                     [[Beta].sub.1]/2[[Beta].sub.2]
                                     [less than] 18.

1c. For proprietary strategists,     ([[Beta].sub.2] +
rates of sales growth will have a    [[Beta].sub.4]) [greater than]
J-shaped relationship with age.      0; and [Mathematical Expression
                                     Omitted], for 15 [less than or
                                     equal to] [[Tau].sub.2] [less
                                     than or equal to] 18, and 0
                                     [less than] [[Tau].sub.1] [less
                                     than or equal to] 4.

1d. Among older firms, age will      ([[Beta].sub.3] [Tau] +
produce higher rates of sales        [[Beta].sub.4] [[Tau].sup.2])
growth in proprietary strategists    [greater than] 0, for 15 [less
than in standards-based              than or equal to] [Tau] [less
strategists.                         than or equal to] 18.

2a. In the years shortly after       ([[Beta].sub.3] [Tau] +
founding, age will produce higher    [[Beta].sub.4] [[Tau].sup.2])
failure rates in proprietary         [greater than] 0, for 0 [less
strategists than standards-based     than] [Tau] [less than or equal
strategists.                         to] 4.

2b. Failure rates will increase      [Delta]r(t) / [Delta] age
with age for proprietary             [greater than] 0, within the
strategists, i.e., those firms       observed range of ages.
will exhibit a liability of          That is, ([[Beta].sub.1] + 2
obsolescence.                        [[Beta].sub.2] [Tau] +
                                     [[Beta].sub.3] +
                                     2[[Beta].sub.4] [Tau]) [greater
                                     than] 0, for 0 [less than]
                                     [Tau][less than or equal to]
                                     18.

2c. Failure rates will generally     [Delta]r(t)/[Delta] age
decrease with age for                [less than] 0, above some
standards-based strategists, i.e.,   threshold, but within
those firms will exhibit a           the observed range of ages.
liability of newness or              That is, ([[Beta].sub.1] +
adolescence                          2[[Beta].sub.2] [Tau]) [less
                                     than] 0, for 0 [less than]
                                     threshold [less than] [Tau]
                                     [less than or equal to] 18.

* Equation 1: r(t) = [[Beta].sub.0] strategy +[[Beta].sub.1] age(t)
+ [[Beta].sub.2] [age.sup.2](t) + [[Beta].sub.3] age(t) [multiplied
by] strategy + [[Beta].sub.4] [age.sup.2](t) [multiplied by]
strategy + [Gamma]X(t).