|
THE COMPETITIVE SEMICONDUCTOR MANUFACTURING HUMAN
RESOURCES PROJECT:
Second Interim Report
CSM-32
Clair Brown, Editor
3. Headcount and
Turnover
Dan Rascher
3.1 Introduction
3.2 Operators
3.3 Technicians
3.4 Engineers
3.5 Temporary
Workers
3.6 Occupation
Ratios
3.7 Conclusion
3.1 Introduction
This section documents the distribution of employment levels and
turnover rates across U.S. and Asian fabs for operators, technicians,
and engineers in order to understand how employment stability,
the use of temporary workers, and the mix of operators, technicians,
and engineers vary across fabs. The employment levels and turnover
rates in these fabs have been affected by external factors such
as the existence of local or national labor markets, cultural
influences, as well as trends in product demand. Poor training
and restructuring were mentioned by a number of the fabs in our
sample as being internal influences affecting employment dynamics.
We find that between 1990 - 1994:
• Employment levels were more stable
in the U.S. than in the Asian fabs.
• The Asian fabs have more turnover (mostly quits) than U.S. fabs.
• Asian fabs use more engineers relative to operators, technicians,
or managers than do U.S. fabs. The ratio of technicians to operators
is fairly constant by product or region.
• A high operator to supervisor ratio, and a high operator to
engineer ratio is negatively correlated with some performance
metrics.
Table 3-1. Employment Levels Across Fabs
| Operators |
ASIA |
U.S. |
| MEMORY |
|
|
| Mean |
653.0 |
- |
| Median |
425.0 |
- |
| Minimum |
386.0 |
- |
| Maximum |
1488.0 |
- |
| Std. Dev. |
471.2 |
- |
| LOGIC |
|
|
| Mean |
382.5 |
342.0 |
| Median |
382.5 |
340.0 |
| Minimum |
379.0 |
257.0 |
| Maximum |
386.0 |
429.0 |
| Std. Dev. |
10.0 |
86.0 |
3.2 Operators
Table 3-1 shows summary statistics
for the fabs by product type and region during the fourth quarter
of 1993. In our sample, employment levels are higher for operators
in the memory fabs than in the logic fabs. Significant regional
differences in operator headcount for logic producing fabs are
not apparent.
Table 3-2. Coefficient of Variation 1990 -
1994
| Operators |
ASIA |
U.S. |
| MEMORY |
|
|
| Mean |
0.14 |
- |
| Median |
0.07 |
- |
| Minimum |
0.05 |
- |
| Maximum |
0.36 |
- |
| Std. Dev. |
0.13 |
- |
| LOGIC |
|
|
| Mean |
0.28 |
0.08 |
| Median |
0.28 |
0.04 |
| Minimum |
0.21 |
0.04 |
| Maximum |
0.36 |
0.15 |
| Std. Dev. |
0.10 |
0.07 |
Employment levels are less stable in
the Asian fabs than in the U.S. fabs over time according to Table
3-2. The Asian fabs have seen a greater increase in employment
over the time period than the U.S. fabs. Part of the increase
is attributable to expansion. Logic producers have experienced
more variability in Asian fabs than memory producers.
Table 3-3. Annual Turnover Information (levels)
| Operators |
ASIA |
U.S. |
| MEMORY |
New Hires |
Quits |
Terminations |
New Hires |
Quits |
Terminations |
| Mean |
114.0 |
113.0 |
13.5 |
- |
- |
- |
| Median |
114.0 |
113.0 |
13.5 |
- |
- |
- |
| Minimum |
57.0 |
78.0 |
0.0 |
- |
- |
- |
| Maximum |
171.0 |
148.0 |
27.0 |
- |
- |
- |
| Std. Dev. |
80.6 |
49.5 |
19.1 |
- |
- |
- |
| LOGIC |
|
|
|
|
|
|
| Mean |
138.5 |
147.5 |
- |
63.0 |
5.0 |
18.0 |
| Median |
138.5 |
147.5 |
- |
63.0 |
5.0 |
18.0 |
| Minimum |
106.0 |
147.0 |
- |
35.0 |
5.0 |
4.0 |
| Maximum |
171.0 |
148.0 |
- |
91.0 |
5.0 |
32.0 |
| Std. Dev. |
46.0 |
0.7 |
- |
39.6 |
0.0 |
19.8 |
The number of new hires is strikingly
close to the number of quits in the Asian fabs, but much lower in
the U.S. fabs. The number of terminations as a percentage of the
number of operators is significantly higher in the U.S. than in
the Asian fabs. However, some of the Asian fabs have a goal of offering
lifetime employment. There is more turnover in general in the Asian
fabs. Some of the Asian fabs have only female operators who work
a very short period of time and then leave the work force to get
married. One fab said that the average amount of time that a female
is with the company is 3.7 years. Some turnover is due to internal
promotions where newly hired operators fill positions left open
by experienced operators who are promoted to the technician job
category.
"Retainment of employees is thought to be the single largest problem,"
reported by one Asian manager. They are faced with a very tight
labor market because of competition from companies who require workers
with similar skills. One company studied why they have such high
turnover rates. They concluded that some of the important reasons
were related to family leave, working the night shift, the desire
to have a service job, and poor training. Another Asian fab even
noticed that a lot of their quits occurred just after major bonuses
were paid.
3.3 Technicians
Table 3-4. Employment Levels Across Fabs
| Technicians |
ASIA |
U.S. |
| MEMORY |
|
|
| Mean |
145.3 |
- |
| Median |
128.0 |
- |
| Minimum |
108.0 |
- |
| Maximum |
200.0 |
- |
| Std. Dev. |
48.4 |
- |
| LOGIC |
|
|
| Mean |
134.0 |
96.0 |
| Median |
134.0 |
73.0 |
| Minimum |
68.0 |
60.0 |
| Maximum |
200.0 |
155.0 |
| Std. Dev. |
93.3 |
51.5 |
Table 3-5. Coefficient of Variation 1990 -
1994
| Technicians |
ASIA |
U.S. |
| MEMORY |
|
|
| Mean |
0.19 |
- |
| Median |
0.24 |
- |
| Minimum |
0.06 |
- |
| Maximum |
0.25 |
- |
| Std. Dev. |
0.11 |
- |
| LOGIC |
|
|
| Mean |
0.17 |
0.14 |
| Median |
0.17 |
0.14 |
| Minimum |
0.10 |
0.12 |
| Maximum |
0.24 |
0.15 |
| Std. Dev. |
0.10 |
0.02 |
Like operators, there are fewer technicians
at the U.S. fabs than at the Asian fabs regardless of product type.
Unlike operators, there is no significant difference in the number
of technicians across product type. Similar to operators, technician
employment levels vary more in the Asian fabs than in the U.S. fabs.
Upon inspection, the number of technicians in the Asian fabs is
continuously fluctuating. The U.S. fabs generally showed mild, steady
increases in technician employment levels.
Table 3-6. Annual Turnover Information (levels)
| Technicians |
ASIA |
U.S. |
| MEMORY |
New Hires |
Quits |
Terminations |
New Hires |
Quits |
Terminations |
| Mean |
14.0 |
7.0 |
- |
- |
- |
- |
| Median |
14.0 |
7.0 |
- |
- |
- |
- |
| Minimum |
14.0 |
7.0 |
- |
- |
- |
- |
| Maximum |
14.0 |
7.0 |
- |
- |
- |
- |
| Std. Dev. |
0.0 |
0.0 |
- |
- |
- |
- |
| LOGIC |
|
|
|
|
|
|
| Mean |
10.0 |
5.5 |
- |
1.0 |
- |
1.0 |
| Median |
10.0 |
5.5 |
- |
1.0 |
- |
1.0 |
| Minimum |
6.0 |
4.0 |
- |
1.0 |
- |
0.0 |
| Maximum |
14.0 |
7.0 |
- |
1.0 |
- |
2.0 |
| Std. Dev. |
5.7 |
2.1 |
- |
0.0 |
- |
1.4 |
As with operators, there is more technician
turnover in the Asian fabs than in the U.S. fabs. The result is
more surprising for technicians because they are mostly male and
continue working even after they get married, unlike the female
operators. This demonstrates that lifetime employment is not a
rigid practice across all of Asia.
Table 3-7. Functional Headcounts - % of total
| Technicians |
ASIA |
U.S. |
| MEMORY |
Process |
Equipment |
Process |
Equipment |
| Mean |
8.6 |
91.4 |
- |
- |
| Median |
8.6 |
91.4 |
- |
- |
| Minimum |
3.2 |
86.0 |
- |
- |
| Maximum |
14.0 |
96.8 |
- |
- |
| Std. Dev. |
7.6 |
7.6 |
- |
- |
| LOGIC |
|
|
|
|
| Mean |
39.5 |
60.5 |
30.0 |
67.0 |
| Median |
39.5 |
60.5 |
20.0 |
80.0 |
| Minimum |
14.0 |
35.0 |
10.0 |
40.0 |
| Maximum |
65.0 |
86.0 |
60.0 |
81.0 |
| Std. Dev. |
36.1 |
36.1 |
26.5 |
23.4 |
Table 3-7 shows the breakdown of the
percentage of technicians by job category across regions for logic
and memory chips. A higher percentage are equipment technicians
rather than process technicians in all product-region combinations.
The ratio of equipment technicians to process technicians is much
higher in memory chip producing fabs than in logic chip producing
fabs. The ratios are similar in a comparison of U.S. and Asian
fabs.
Go to Next Section of this Chapter
Go to Table of Contents for this Chapter
Go to Table of Contents for the CSM-HR Interim Report
|