Full text of New England Economic Review : July 1966

View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

NE\N ENGL AND

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

Does Retraining Pay?
This article evaluates the costs and benefits of
public retraining programs. It concludes that
the benefits are particularly high for society as
a whole, thus justifying government spending.

New England 's Gross Product
Estimates of the region's gross product indicate
that it doubled during the 1950-1964 period.
Among the states, Connecticut led in relative
growth, followed by New Hampshire.

Does Retraining Pay?
Nation's over-all unemployment rate
has declined this year to the lowest level
since 1953. It is rapidly approaching the level
of a fully-employed economy. Nevertheless,
certain members of the labor force continue to
experience relatively high levels of unemployment. For example, the unemployment rate
both for teenagers and for unskilled workers
continues around 10 percent. Moreover, about
one-fifth of the unemployed have been without
work for 15 weeks or more.

At the same time the demand for highlyskilled workers continues strong and exceeds
the supply in some areas. Such shortages of
labor can slow up the economy's expansion and
put pressure on costs and prices. The unemployed all too often do not have the skills
required to fill the available jobs.
Thus, the need continues for programs designed to upgrade the skills of certain workers
and to retrain some workers in new skills. Such
programs are functioning and expanding
rapidly. In 1965 more than 100,000 persons
completed training under the Manpower Development and Training Act. Three out of four
2

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

of these individuals were placed in jobs within
90 days after completion of training. This year
it is anticipated that the number trained will be
about twice that of last year.
To evaluate these public training programs it
is necessary to determine the costs and benefits
derived from_ them. Costs are incurred and
benefits are received by the worker, the
government, and the economy as a whole. The
important questions are who receives the
greatest benefits in relation to costs and therefore who should pay for the training. At
present, most training is paid for by private
employers. 1 On the other hand, because the
benefits of these programs are greater for
society as a whole than they are for private
firms, not all retraining can be left to private
initiative.
Michael Borus, while a doctoral candidate at
Yale University, studied the costs and benefits
of a retraining program conducted in the early
1960's in Connecticut to help resolve these
questions. He concluded that the benefits from
1Edwin F. Estle, "Industrial Investment in Manpower," New
England Business Review, Federal Reserve Bank of Boston, February,
1964.

July 1966
government-sponsored retrammg outweighed
the cost for all groups involved: the worker, the
government, and the economy as a whole.
Retrained workers who used their new skills
earned higher incomes and suffered less unemployment than they would have without retraining. The extent to which benefits exceeded
costs varied, however, with the worker's characteristics. Probably because of their greater
adaptability and motivation, the younger,
more educated workers and those who were
employed at the time they applied for retraining were more likely to find jobs requiring their
new skills. However, when unemployed workers did succeed in finding jobs after being retrained they received the greatest benefits. For
all groups, workers' benefits exceeded their own
costs when they participated in retraining programs sponsored by the government.
On the other hand, many fewer workers
would enter retraining if they rather than the
government were forced to pay the full bill. If
the worker were to pay all the cost of retraining,
his benefit-cost ratio would fall to low levels. In
some cases, the worker's costs might exceed his
benefits. Yet for society as a whole the benefits
of retraining usualJy greatly exceed the costs
thus justifying government spending for these
programs.
Likewise, society would lose if all retraining
were left to private firms. The present government programs operate where labor shortages
have existed for long periods. This indicates
that the firms consider the retraining unprofitable due to the risk that the worker would leave
the firm after completing training. Nevertheless, the gains to the economy from retraining would be significant.

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

Borus concludes, therefore, that the government should be willing to carry the costs of
many retraining programs. His study also suggests, however, that the government retraining
programs are not a complete answer in solving
a chronic unemployment situation. First of all,
only '/3 percent of those eligible to take the
course actually enrolled. Of those who did not
enroll, one-third continued to remain unemployed. Moreover, only two-thirds of those
who enrolled in the courses made use of their
training. The remainder of the workers either
withdrew from the courses (20 percent) or
graduated and found jobs which did not require
the retraining (13 percent). In general, the
older, the less educated, and the long-term unemployed worker had less chance of using and
benefiting from the retraining.
This indicates that supplementary programs
continue to be needed to raise the educational
level of adults as well as youngsters, to provide
guidance and counseling services, and to rmprove information channels on existing job
opportunities.

The State of Connecticut was one of the first
in the country to offer publicly-supported
classes specifically designed to retrain unemployed workers. These early efforts permitted
the selection of a sample of workers with
sufficiently long post-retraining employment
records to judge the impact of retraining. A
personal interview survey was made of more
than 300 male workers who had qualified for
retraining courses in Connecticut. Some of the
workers did not take part in the retraining
programs; others did not complete the retraining course they started.

New England Business Review
The study group was selected from male
workers eligible for courses given between May
1961 and March 1962 in machine shop operations, ship.fitting, and pipe.fitting. Since the
effects of retraining may depend on the level of
unemployment- that is, workers may be
more easily placed in areas of higher employment - four different labor markets were
chosen with unemployment rates varying from
3 to 12 percent. The men in the study group
were considerably younger than most of the
unemployed. The median age of the men
studied was 23 years compared with 39 years
for the unemployed labor force. As a result of
their youth, more than half the men in the
sample were unmarried, and about three-tenths
were just entering the labor force. The experienced workers in the group were mostly unskilled or semiskilled.
Not only were the men in the sample younger,
they were also more educated than the average
male population of Connecticut, and considerable more than the unemployed. Only a third
of the men eligible for retraining had not
graduated from high school in contrast with 60
percent of the men in the State.
About one-third of the sample group were
employed at the time they took the aptitude
test for retraining. Of those workers who were
unemployed, about one-fourth had been unemployed six months or longer.

To measure the benefits of retraining, the
experience of those who completed the retraining course was compared with those who didn't
start, those who dropped out before completion,
and those who completed retraining but didn't
use their new skills. Because all the workers
involved in the study met the minimum qualifi4

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

cations necessary for retrammg, the major
differences in income and employment could be
assumed to reflect the effects of retraining.
The study found that, on the average, the
worker who utilized his retraining increased his
wage income by $500 a year and reduced his
annual unemployment by about 5 weeks in
the first year after retraining. These are not,
however, the net benefits received by the
worker. Some of the increased income would go
into income and Social Security taxes. Also an
allowance must be made for the reduction in
unemployment compensation benefits received
by the worker who utilized retraining as compared to what would have been received if he
had not been retrained. His net income mcreased in the first year by only $300.
The individual worker's costs of retraining
were largely the transportation costs to and
from classes and the cost of income lost during
the course. The average transportation cost per
retrained worker in the sample was $25. The
cost of lost income is dependent upon the
worker's expected employment status and earnings if he does not enter the course. If the
worker expects to be unemployed during the
retraining course, he will actually increase his
income by receiving training allowances greater
than his normal unemployment benefits. On
the other hand, if the worker expects to be
employed during the period of the course if he
does not take training, he will experience a
reduction in income by the amount to which his
earnmgs would have exceeded training allowances.
To determine these costs it becomes necessary
to make several assumptions: for example,
assumptions as to what the retrainee could have
earned if not taking retraining, what the tax

July 1966
rate on his wage income would be, which in turn
depends upon the number of dependents, and
what the difference in transportation costs
would be if he were employed.
This points up the difficulty in making bene fit-cost comparisons. Some subjective judgments must enter into the calculations since
expectations are involved. It is not possible to
give one benefit-cost ratio, but only to give the
possible range of ratios depending upon the
assumptions made .
The benefits for the worker from retraining,
of course, extend beyond the first year if he continues in the new job. Assumptions must again
be made as to the probability of his staying about 18 percent of the sample group left their
new jobs for occupations unrelated to retraining
within 12 months after graduation - and the
appropriate discount r ate to apply to future
benefits to determine their present value.
On the basis of the sample data and h is set of
assumptions, Borus concluded that the benefitcost ratio for the average worker would range
between 3 and 6. That is, for every dollar of
cost to the worker for retraining, he would re ceive $3 to $6 in benefits. If the retrainee was
among the long-term unemployed, however, the
benefit-cost ratio would be somewhat higher.
For although he will have a lower probability of
completing the course and finding a trainingrelated job, the worker unemployed for more
than 15 weeks who used his training earned
average benefits approximately twice as great as
the worker who was unemployed for a shorter
period. Borus attributes these higher benefits to
the scarcity of alternative work opportunities
for the long-term unemployed if they are not
retrained.

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

This article is based on a Yale University
doctoral dissertation by Michael E. Borus, now
Assistant Professor at Michigan State University. Copies of the complete report are available from the Research Department on request.

Retraining costs for the government include
those of selecting and processing trainee applicants, paying retraining allowances, paying for
instructors, tools, etc. Benefits for the government would be principally the reduction in costs
of unemployment and public assistance and the
increase in tax receipts which resulted from the
higher income levels achieved through retraining. These benefits would again depend
upon the characteristics of the trainees and
their likelihood of leaving their new jobs.
Borus concludes that benefits would exceed
costs in the range of 11 to 40 times for the
government. These ratios would decline
slightly, however, if the number of weeks of
retraining were lengthened (the average length
of the courses studied was only 6 weeks), and if
training allowances were increased. The ratios
would decline drastically if retrainees merely
repJace other employed workers rather than add
to total employment. Similarly, the ratios
would fall sharply if retraining caused a shift
in spending patterns that resulted in reduced
employment in some other part of the economy.
For example, firms employing retrained workers
generally increase their output. If as a result
more of their products are purchased in lieu of
the products of other firms, employment in these
other firms would be reduced. Thus, the reduction in unemployment compensation and public
assistance in communities offering retraining
programs might he largely offset because of increased unemployment in other places.
,)

New England Business Review
The costs to the economy are much the same
as those to the government except for the retraining allowances which constitute a cost for
government but are merely a redistribution of
income for the economy as a whole. Benefits
would depend upon the same factors as for the
workers. However, they would be larger because of the assumption that the total value of
the worker's production is to be included,
rather than just the increase in his output.
Borus' assumptions for the economy as a
whole are based on the premise that the jobs for
which retraining was given would remain unfilled if retraining did not take place, and that
the economy is at a less than fully-employed
level. Thus, the entire output of the retrained
worker becomes a benefit to the economy.

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

The economy's benefit will fall as the full employment level is reached. The retrainee will
move from a lower to higher skill job, but there
will be no one unemployed to take the lower
skill job. In this instance benefits are raised
only by the increase, not the full amount of the
retrainee's output. In addition, even in a less
than fully-employed economy it may be that
the jobs would not remain unfilled if retraining
were not conducted. For example, firms might
be able to redefine jobs in such a way as to
employ more unskilled or semi-skilled wqrkers.
Or again, firms might extend their normal work
weeks to obtain the increased output.
Even with a more restrictive set of assumptions, however, the benefits to the economy
from retraining would b~ well above the costs.

July 1966

New England's Gross Product
total output of goods and services in the
United States, the Gross National Product,
in 1965 was $676.3 billion. This estimate, and
it admittedly is only an estimate, was made by
the United States Department of Commerce.
It is subject to some revisions as additional information becomes available. Even with these
qualifications, however, it is the best overall
measure of the national economy that is available. It is one of the major guides to economic
policy formulation and business planning.

For individual states and regions of the
Nation there is no official estimate of the total
output. The most complete measure now available is the personal income series published by
the U. S. Department of Commerce. This is
quite useful for measuring economic performance. Nevertheless, it provides information
only on the amount of income received in an
area, it does not measure the value of goods
and services produced in an area.
A gross product measure at the state and
regional levels provides a framework for making
economic projections by industry that can be
The New England Business Review is produced
in the Research Department. Edwin F. Estle
was primarily responsible for the article, "New
England's Gross Product."

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

related to gross national product projections.
These state and regional projections can be of
use in the planning and policy formulation
activities of government and business in the
region.
The Federal Reserve Bank of Boston has,
using the personal income data and other information, estimated New England's gross
product for the 1950-1964 period. The estimate
indicates that total output in current prices
doubled, a 100 percent increase, over the 14year period to a total of $41.5 billion in 1964.
The ation's gross product shows a somewhat
greater increase, rising 121 percent from 1950
to 1964. Real output, that is gross product
adjusted for price changes, in New England
advanced by almost half, 47 percent, over the
14-year period in terms of 1958 dollars. Thus,
the average annual growth of gross product in
the region over the period was 2.8 percent.
The Nation's average annual growth in real
output over the period was 3.5 percent.
The region's personal income expanded, in
current prices, by 109 percent over the 14
years, or at a rate 9 percent faster than gross
regional product. Consequently, the personal
income level in relation to gross product rose
from 73 percent in 1950 to 76 percent in 1964.
In the intervening years, however, the ratio

New England Business Review
varied considerably. In 1956 it was 84 percent,
while in 1959 it was down to 74 percent. The
distinction between personal income and gross
product is rather complicated. To arrive at
personal income, it is necessary to deduct certain items from gross product and to add others.
The deductions include depreciation, indirect
business taxes, and undistributed profits. The
additions are primarily transfer payments to
persons from business and government, e.g.
benefits from social insurance funds.
Personal income, therefore, does not represent all income (e.g. undistributed profits) or
economic activity of an area. It deviates significantly from gross product over the business
cycle. In recession years personal income does
not decline, relatively, as much as gross product
because of the anticyclical effect of transfer
payments. In years of high business activity
it advances relatively less because of the exclusion from personal income of such things as
undistributed profits which advance more
rapidly in such periods. To illustrate, Chart 1
shows that gross product advanced 5 percent
from 1956's level in 1957, a peak of the business
cycle in New England. This was almost double
the rate recorded for personal income. A comparison of the levels in 1958 with those of 1957
shows the divergence in a recession period.
Gross product declined by 2 percent, while
personal incomes rose by that percentage.
S

The method used to estimate the region's
gross product provides a distribution of the
total by sector and industry. Gross product
from farming was found to he $390 million in
1964, a decline of one-tenth from its level in
1950. Another industry which showed a decline was railroads, where gross product fell
8

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

Chart 1
NEW ENGLAND'S GROSS PRODUCT
AND PERSONAL INCOME

Billions of Dollars

Gross Product
30

Personal Income

or 1 I I I
1950 1952 1954

1956 1958

1960

I If
1962 1964

$37 million, or 15 percent, over the 14-year
span. All other types of transportation, on the
other hand, registered an advance in gross
product of more than four-fifths during the
period.
As the accompanying table shows, the greatest relative increases in output occurred in the
communications-public utilities and government sectors, where output increased almost
one and three-fourths times. Other sectors
where output more than doubled were services,
finance-insurance-real estate, and construction.
Manufacturing, which accounts for almost a
third of all output from the region, increased

July 1966
its output by four-fifths in the 1950- 1964
period.
In summary, the nonmanufacturing industries with the exception of farming and railroads
have been expanding their output, primarily
in the form of services at a rapid rate. On the
other hand, manufacturing has shown a slower,
hut still sizeable, growth in product. Thus, it is
clear that New England's output of services has
risen faster than its output of goods.

Gross product estimates were also prepared
for each New England State. As shown in the
accompanying table, Connecticut led the six
States with a gross product increase of 128
percent over the 1950- 1964 period. This was
a greater relative advance than in the Nation
as a whole. New Hampshire was in second
place, with a gain of 112 percent. Almost twothirds of this rise occurred in the 1958-1964

NEW ENGLAND'S GROSS REGIONAL PRODUCT
1964

$ Million

Economic Sector

Private Nonfarm
Mining .....
Construction
Manufacturing
Trade . . . . .
Finance, Insurance and Real Estate .
Transportation . . . . . . . . . . . .
Communications and Public Utilities ..
Services ..
Government
Farm ...
Total

Percent Change
from 1950

38,313
90
2,092
13,344
7,534
5,964
1,023
1,804
6,460
2,801
390

+ 98%
+ 96
+116
+ 78
+ 72
+133
+ 49
+173
+158
+172
- 10

41,504

+100

1964

$ Million

Percent Change
from 1950

12,589
2,508
20,400
1,947
2,816
1,148

+128%
77
+ 92
+112
66
92

GROSS PRODUCT BY STATE
State

Connecticut .
Maine . . . .
Massachusetts
New Hampshire
Rhode Island
Vermont . . . .

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

+
+

New England Business Review
period. Rhode Island trailed the six States in
gross ·product growth, as it did in personal income growth, primarily because of the slow
growth in manufacturing output. Gross product from manufacturing in this State grew only
by slightly more than a fourth over the 14-year
period. Durable goods industries in Rhode
Island had a 70 percent growth in output,
whereas nondurable goods industries registered
an output increase of only 6 percent, largely
because of the decline of textiles, in the 19501964 period.

Relation of Gross Product to Other
res
The gross product estimates can be used in
conjunction with other measures such as population, capital investment, and employment to
give a more detailed picture of the regional
economy.

Manufacturing' s Investment and Gross Product
Another comparison which can be made is the
amount of investment which manufacturers
have made in relation to their output. The
Bank has conducted surveys of New England
manufacturers' capital expenditures since 1957.
As a proportion of manufacturers' gross output,
these outlays have been rather stable. In 1957
they amounted to 7 percent of gross output.
Since then, however, they have held at 5 percent of gross output.
National comparisons show much the same
pattern. In 1957 capital outlays were 12 percent of gross product in manufacturing. In
the years following they were at 9 to 10 percent
of gross product.

Real Product Per Capita
From 1950 to 1964 real product per capita
in New England increased a fourth, going from

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

Chart 2
NEW ENGLAND'S GROSS PRODUCT
BY MAJOR INDUSTRY

Billions of Dollars
16
14

Manufacturing
12
10

8
6
4

Services
Construction

2L---------------0.__.......____.____._......._......._____._......__.____.__.__.......____.______....___.

1950 1952

1954

1956

1958 1960

1962

1964

$2,779 in 1950 to $3,458 in 1964. The 1950
level was 17 percent above the national average.
In 1964 however, the region's per capita
product relative to the Nation had declined,
being 14 percent above the national level of
$3,019.

Real Product Per Manhour
Using the gross product estimates for manufacturing, converted to real product by price
deflation, it is possible to compute the real
product per manhour in manufacturing. In
terms of the 1957-1959 level, output per manhour was 87 .5 percent in 1950. In 1963 the
index was 114.1. These index levels are very
close to those for the Nation's manufacturing

July 1966
sector. Nationally, _output per manhour was
85.1, with 1957-1959 equal to 100, in 1950 and
115.4 in 1963. Thus, the movement of productivity in the region's manufacturing sector
appears quite similar to that for the Nation as
a whole. However, the level of output per man-

hour is lower in New England than in the
Nation because of the region's concentration
in labor-intensive, rather than capital-intensive, industries. In 1962, for instance, output
per manhour in New England was about onetenth below that of the Nation.

Validity of the Estimates
Not quite three-fifths of the total gross product was obtainable from published sources,
the remaining two-fifths was estimated by the
Bank. The estimates involve an assumption
that the ratio of personal income originating in
current production to gross national product in
the Nation applies also to New England. It is
an assumption that the structure within each
industry group in New England is similar to that
of the Nation. This is certainly not completely
true, but the errors involved may be offsetting.
There also may be some upward bias in the
New England estimates in regard to capital
consumption allowances by assuming a structure similar to that of the Nation. New England
producers tend to use relatively less capital
equipment than does industry nationally. However, the region's growth relative to the Nation
is about the same for net product, where capital
consumption allowances are excluded, as for
gross product. Thus, the movement of output
over time was not affected by the possible
upward bias.
The procedure does take account of differences in the major-industry structure between
the New England and national nonfarm private
sectors. It does adjust for the region's relatively
greater concentration in manufacturing and
relatively less dependence upon the other
sectors.
Some checks on the estimates are possible.
One that can be made on the manufacturing
gross product estimate is to compare it as an
index, deflated for price changes, with the
Bank's index of manufacturing production.

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

The comparison shows them to move together,
and to deviate by no more than 4 percentage
points in any given year. In 1964, for example,
the deflated index of manufacturing gross output was 118 in terms of 1957-1959 equalling
100. The Bank's manufacturing production
index was 122 in that year. Given certain
differences in concept, methods of aggregation,
etc., the divergence of the two indexes is largely
explained. Moreover, the comparison at the
national level of real product and industrial
production shows much the same type of divergences. In addition, the gross product estimates tend to follow fairly closely the movement
of value-added data from the Census of Manufacturing and Mining. It does seem, therefore,
that the gross product estimates presented
here are reasonable approximations of the
region's output. 1
There is no direct check upon the gross
regional product estimates. However, the
movement of the ratios derived - gross product
per capita, manufacturing investment to gross
product, and gross output per man hour - lend
plausibility to the estimates. These ratios move
much like their national counterparts and are at
levels consistent with other measures of the
regional economy.
1The

method of estimating New England's gross
product is essentially that described by John W.
Kendrick and C. Milton Jaycox in "The Concept and
Estimation of Gross State Product," The Southern
Economic Journal, (October 1965) pp. 153-168.
A technical supplement is available from the Bank,
giving the methods used and the resulting
estimates.

July 1966

Here's New England NEW ENGLAND
MANUFACTURING INDEXES (seasonally adjusted)
pMay '66 Apr. '66
May '65
1957-59 = 100

UNITED STATES
May '66
Apr. '66
May '65

All Manufacturing

146

132

158

156

143

Nonelectrical Machinery
Electrical Machinery
Transportation Equipment

170
171
187

165
171
185

144
145
163

178
185
166

174
184
166

157
157
147

Textiles, Apparel, Leather

118
117
116
113

119
120
117
113

108
107
114
104

142
142
150
114

142
142
150
116

135
132
145
111

133

131

128

153

150

141

Textiles
Apparel
Leather and Shoes
Paper

Percent Change from:

May'66
2,435

Ap~'66
+ 2

May'65
+21

May '66
53,311

Deposits ($ millions)
(Weekly Reporting Member Banks)

6,540

+11

164,053

Check Payments ($ billions)
(Selected Metropolitan Areas)*

222.4

+22

3,348.1

Consumer Installment Credit Outstanding
(index, seas. -adj. 195 7-59 = 100)

171.0

BANKING AND CREDIT
Commercial and Industrial Loans($ millions)
(Weekly Reporting Member Banks)

DEPARTMENT STORE SALES
(index, seas. adj. 1957-59 = 100)

EMPLOYMENT, PRICES, MAN-HOURS
& EARNINGS
Nonagricultural Employment (thousands)
Insured Unemployment (thousands)
(excl. R.R. and temporary programs)
Consumer Prices
(index, 1957-59 = 100)
Production-Worker Man-Hou rs
(index, 1957-59 = 100)
Weekly Earnings in Manufacturing($)

127

4,144
66

1
-20

114.1
(Mass.)

208.7

Apr. '66
+ 1

May '65
+18

+17

+12

n.a.

+ 4
-30

63,070
885

+ 1
-18

112.6

n.a .

-29

106.6

116.1

104.70
(Mass.)

112.05

OTHER INDICATORS

Total Construction Contract Awards** ($ thous.)

304,912

+11

+27

4,988,758

+14

Residential

115,584

+15

+12

2,018,352

+13

Nonresidential

124,851

+57

1,811 ,676

+12

+16

64,477
166

+11
+10

1,158,730

-17

968

Public Works and Utilities
Electrical Energy Production (4 weeks
ending May 21st 1966)
(index, seas. adj. 195 7-59 = 100)
Business Failures (number)
New Business Incorporations (number)
•Seasonally adjusted annual rate.
**3-mos. moving averages -Mar., April, May

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

6
+12

p = preliminary

+22

+14

175

997

-10

-16

17,036

n.a. = not available

Full text of New England Economic Review : July 1966

FRASER