John Carroll University

Haight, Alan Day

Carroll, Michael C.

The Journal of Economics and 
Politics 
Manuscript 1162 
DOLLAR APPRECIATION AND LAYOFFS IN OHIO 
MANUFACTURING: TIME SERIES EVIDENCE 
Alan Day Haight 
Michael C. Carroll 
Follow this and additional works at: https://collected.jcu.edu/jep 
Journal of Economics and Politics, Vol. 16, No. 1, 2003 - 2004. 
DOLLAR APPRECIATION AND LAYOFFS IN OHIO 
MANUFACTURING: TIME SERIES EVIDENCE 
ALAN DAY HAIGHT 
ASSOCIATE PROFESSOR OF ECONOMICS 
BOWLING GREEN STATE UNIVERSITY 
MICHAEL C. CARROLL 
ASSISTANT PROFESSOR OF ECONOMICS 
BOWLING GREEN STATE UNIVERSITY 
ABSTRACT 
IT SEEMS OBVIOUS THAT APPRECIATION OF THE DOLLAR MUST DAMAGE THE COMPETITIVENESS OF OHIO MANUFACTURERS, 
BUT ECONOMETRIC EVIDENCE SUGGESTS A SMALLER EFFECT THAN ONE MIGHT EXPECT. EVIDENCE FOR THE PERIOD 1973 TO 
2002 INDICATES THAT A FOURTEEN PERCENT APPRECIATION OF THE DOLLAR LEADS TO A ONE PERCENT (I.E., TEN THOUSAND 
JOB) 
REDUCTION 
IN OHIO'S MANUFACTURING EMPLOYMENT BASE. THIS BY ITSELF CANNOT EXPLAI  THE STATE'S NET LOSS 
OF 
NEARLY 
400 THOUSAND MANUFACTURING JOBS.SINCE 1978; IN FACT IT WOULD TAKE A 38 PERCENT DEPRECIATION OF THE 
DOLLAR TO COUNTERACT JUST ONE YEAR'S SE'CULAR DECLINE IN THE STATE'S MANUFACTURING EMPLOYMENT. 
MANUFACTURING SECTORS THAT SELL STANDARDIZED COMMODITIES (SUCH AS LUMBER, GLASS, AND PRIMARY METALS) ARE 
SOMEWHAT MORE SENSITIVE TO THE EXCHANGE RA TE THAN SECTORS THAT PRODUCE MORE HEAVILY PROCESSED 
COMMODITIES. THE LATTER SERVE SPECIALIZED NICHE LOCAL MARKETS AND ARE PARTIALLY HEDGED AGAINST DOLLAR 
FLUCTUATIONS BY THEIR PURCHASES OF TRADABLE INPUTS. 
I. INTRODUCTION 
The big run-up of the dollar in the early 1980s coincided with a big run-down of Ohio's 
manufacturing employment. In the late 1990s the dollar was again on the rise (indeed, by 2002 the 
dollar was nearly a third stronger than it had been seven years earlier) and Ohio's manufacturing 
employment was falling again; so in the late '90s as in the early '80s there appeared to be good reasons 
for blaming Ohio's manufacturing woes on the strong dollar2°. The objective of this paper is to test that 
possibility more formally. 
As in so much of macroeconomics, the effect of a dollar appreciation upon employment 
2°For recent studies of the Ohio employment picture, see Cassell and Hanauer (2000). 
At the national level, there is a continuing debate on this issue. See for example Sachs and 
Shatz (1994), Brenner (2002), and Goldberg and Knetter (1997). 
113 
----· 
1
Haight and Carroll: DOLLAR APPRECIATION AND LAYOFFS IN OHIO MANUFACTURING: TIME SERIE
Published by Carroll Collected,
Journal of Economics and Politics, Vol. 16, No. 1, 2003 - 2004. 
depends upon whether wages are sticky or flexible . For example, a rise in the nominal Yen/$ exchange 
rate tends to increase the Yen price of the Jeeps made in Toledo. Jeep will reduce its dollar price to 
dampen the rise in the yen price and thereby prevent a loss of sales volume overseas; but the lower 
dollar price effectively reduces the marginal revenue product of labor. That fall in Jeep's derived 
demand for labor will reduce wages in the classical case where wages are flexible, but it will reduce 
employment in the Keynesian case where wages are rigid. The proposition to be tested here is that 
Ohio's manufacturers face sticky wages, so that a dollar appreciation causes layoffs. 
II. ECONOMETRIC EVIDENCE FOR OHIO MANUFACTURING 
Using data for the period 1973 to 2002, regressions were run to estimate the following 
equation, representing the derived demand for Ohio's manufacturing workers: 
OHManujEmp= A* (realXR /J)(WorldCondtionsr) ( eq. 1) 
where "realXR" is the real exchange rate (with a rise indicating a real appreciation of the dollar) and 
"WorldConditions" is a control variable indicating the industrial output of the world's industrial 
countries. Our main interest is in the elasticity of employment with respect to the real exchange rate; 
we expect this elasticity to be negative and significant if there are layoffs after dollar appreciation.21 
Monthly data on Ohio manufacturing employment since 1973 (when the dollar began floating 
after the collapse of Bretton Woods) was obtained from the Bureau of Labor Statistics. Data on the real 
exchange rate was obtained from the Federal Reserve (FREDII). For data on world manufacturing 
conditions, a time series from the International Monetary Fund (industrial production by industrial 
21 In practice it is difficult to tell whether a change in the dollar's real exchange value 
represents a short-run overvaluation (which would cause layoffs) or an equilibrium change in 
the real exchange rate (which might not). At the IMF, Bartolini ( 1995, p.4 7) noted that "while 
it is useful to think of 'competitiveness' as the relative position of a country's exchange rate 
with respect to its long-run equilibrium value, it is not always easy to distinguish empirically 
between a change in the real exchange rate that reflects a change in competitiveness or a 
structural change that leads to a persistent change in the equilibrium real exchange rate." 
Equilibrium changes would not typically be reversed in a few years, however, so the 49 percent 
run-up AND then run-down of the real exchange rate during the 1980's remains hard to explain 
with real factors. The past twenty-five years have produced two big spikes (bubbles?) in the 
real exchange rate, suggesting that short-run overvaluation does occur. 
114 
2
Submission to The Journal of Economics and Politics
https://collected.jcu.edu/jep
DOI: 10.59604/1046-2309.1162
Journal of Economics and Politics, Vol. 16, No. I, 2003 - 2004. 
countries) was used. 
Initially a regression in terms of the levels of the logs of the variables was used (controlling for 
seasonality with I I dummy variables in the traditional manner). Those preliminary results indicated an 
elasticity of Ohio manufacturing jobs with respect to the real exchange rate of minus one-sixth; but 
(despite an impressive-looking t-value of -5.7) this result could not be trusted because a Dickey-Fuller 
test of the residuals gave a tau-statistic of only -1.56 (p=.5 I), which isn't sufficiently negative to reject a 
unit root and show stationarity of the residuals.22 
To reduce the spurious correlation problems involved with nonstationary time series, the data 
was differenced. After differencing, equation 1 becomes 
Mog(manufEmp) = Mog(A) + fJMog(realXR) + yMog(WorldCondtions) 
(eq.2) 
where the changes in logs represent annual percentage changes, which are less likely to contain 
stochastic trends than the levels of those variables. Using OLS, the 12-span change in the log of Ohio 
manufacturing employment was regressed on the 12-span changes in the real exchange rate and world 
industrial conditions (with an intercept term to capture any drift). This gave the following results (t-
values in parentheses): 
/1 12 LogManujEmploy = -.027 -.07 !1 12 Log_ realXR + .87 !1 12 LogWorldlnd _prod (eq. 3) 
(-17) (-2.8) (23.4) 
22 The critical values for the Dickey-Fuller tau-statistic are more stringent than for the 
traditional t-test. (This test used a mean with no trend.) 
115 
--- - -- - ~ ----- ---
3
Haight and Carroll: DOLLAR APPRECIATION AND LAYOFFS IN OHIO MANUFACTURING: TIME SERIE
Published by Carroll Collected,
Journal of Economics and Politics, Vol. 16, No. 1, 2003 -2004. 
A Dickey-Fuller test of the residuals gives a tau statistic of -3.64 which is significant (with an intercept, 
for 341 observations) at the .OJ level.23 
The intercept term in equation 3 (which represents the change in the log of A in equation 2) 
indicates that Ohio has been losing 2. 7 percent of its manufacturing jobs per year, as a long term secular 
trend. The estimated exchange rate elasticity is minus .07, or roughly one-fourteenth. (Note that this is 
smaller than the one-sixth that was estimated from the regression in terms of levels). This suggests that 
it would take a fourteen percent appreciation of the dollar (in real terms) to cause a loss of one percent 
of Ohio's manufacturing jobs. Another way to think of this is that it would take a 38 percent 
depreciation of the dollar to counteract just one year's secular decline in manufacturing employment. 
III. RESULTS FOR DETAILED MANUFACTURING SECTORS 
Equation 2 was also estimated for more detailed sectors within manufacturing. Table 1 shows, for each 
sector, the secular trend and the estimated exchange rate elasticity. Values of the t-statistic that are 
insignificant at the 10% level are shown in parentheses. [Also shown for each sector is the probability 
associated with the Dickey-Fuller tau-statistic testing for stationarity of the residuals; for sectors where 
this probability is less than 0.10, the null hypothesis of a unit root in the residuals can be rejected; this 
rejection indicates that the residuals are stationary as desired.] 
Table 1 
Elasticity of Ohio employment with respect to the real exchange rate (XR) 
OH Manuf. Sector 
Lumber and Wood 
Stone Clay & Glass 
Primary Metals 
Secular 
Trend 
XR elasticity t-statistic 
onXR 
Pr(unit root) 
in residuals 
.002 
-.17 
-3. l 
.003 
-.03 
-.11 
-3.8 
.0001 
-.05 
-.18 
-3.9 
23That is, there is a .01 probability that there is a unit root in the residuals; this is small 
enough to reject the null hypothesis of a unit root, indicating that the residuals are stationary. 
116 
4
Submission to The Journal of Economics and Politics
https://collected.jcu.edu/jep
DOI: 10.59604/1046-2309.1162
Journal of Economics and Politics, Vol. 16, No. l, 2003- 2004 . 
. 004 
Fabricated Metals -.03 
-.05 
(-1.5) .004 
Industrial Machinery -.03 
-.07 
-1.7 
.09 
Electronic and electrical .00 .00 
(-
0.0) .005 
Instruments -.01 -.01 
(-0.2) .001 
Transportation equipment -.03 -.03 (-0.6) 
.000 
Food -.07 
-.07 -4.l 
.000 
Textiles -.04 .02 
(0.5) .001 
Paper 
-.02 -.02 
(-
0.7) .000 
Publishing -.12 
-.12 
7.3 .001 
Chemicals 
.00 
-.14 
7.4 
.006 
Rubber 
-.02 
.00 
(0.1) 
.000 
All OH Manufacturing -.027 -.07 
-2.8 
.006 
117 
- -- - -----~---------
5
Haight and Carroll: DOLLAR APPRECIATION AND LAYOFFS IN OHIO MANUFACTURING: TIME SERIE
Published by Carroll Collected,
Journal of Economics and Politics, Vol. 16, No. I, 2003 -2004. 
For example, in the stone clay and glass sector there is a long-run tendency to lose 3 percent of the 
jobs per year, and a 9 percent depreciation of the dollar would save 1 % of that sector's jobs. The table 
shows that sensitivity to the exchange rate varies widely by sector. As a general rule, it appears that the 
extractive and primary sectors (such as lumber, glass, and primary metals) are more sensitive to the 
exchange rate than the more "downstream" sectors (such as fabricated metals, transportation equipment, 
electronics & electrical, instruments, paper, textiles & apparel, or rubber).24 This may be in part 
because the more advanced stages of processing often involve buying importable inputs: a strong dollar 
lowers the cost of these items, improving competitiveness. In effect, downstream processes that buy 
imported inputs (such as instruments and electronics) are to some extent hedged against a strong dollar. 
Another explanation for reduced sensitivity downstream may involve product differentiation. For 
example, employment in primary metals is significantly affected by the exchange rate, yet employment 
in fabricated metals is not significantly affected. This may reflect the fact that primary metals are often 
more standardized (and hence more tradable) commodities than are fabricated metals; the latter may 
gain some protection from foreign competitors (and hence form exchange rate fluctuations) by serving 
specialized domestic "niche" markets. 
However, exceptions do occur to the pattern of reduced exchange rate vulnerability 
downstream. For example, employment in publishing is apparently more sensitive to the exchange rate 
than employment in the paper industry; perhaps publishing is more "footloose" internationally. 
IV. CONCLUSION 
This paper looks at Ohio manufacturing employment since 1973, when the dollar began to float 
on foreign exchange markets. Certainly more research is needed because the effect of a change in the 
dollar's foreign exchange value varies depending on a variety of factors, including price flexibility and 
unemployment (Bartolini, 1995). The tentative conclusion reached here, however, is that "common 
sense" (which is perhaps based upon intuitive experience with the levels of the variables) probably 
overstates the strength of the connection between the dollar and Ohio's manufacturing employment. 
Spurious correlation can give the impression of a stronger connection than actually exists. A more 
realistic picture is obtained by la regression based on the percentage changes in the variables; this 
indicates only a weaker (although still statistically significant) association. 
24See Rennie (1997) for a discussion of the regional implications. 
118 
----------------------------------------1111 
6
Submission to The Journal of Economics and Politics
https://collected.jcu.edu/jep
DOI: 10.59604/1046-2309.1162
Journal of Economics and Politics, Vol. 16, No. 1, 2003- 2004. 
The sensitivity of employment to exchange rate fluctuations varies widely across manufacturing 
sectors. In general, sectors that sell standardized commodities (such as lumber, glass, and primary 
metals) are more sensitive to the exchange rate than sectors that produce more heavily processed 
commodities. The latter sometimes serve specialized niche local markets and are partially hedged 
against dollar fluctuations by their purchases of imported inputs. 
Our overall conclusion about the effect of a strong dollar on Ohio manufacturing employment 
must be treated as quite approximate, because further research is needed.25 Our research so far 
indicates that it takes a fourteen percent appreciation of the dollar to induce a one percent reduction in 
Ohio's manufacturing employment. For example, if the dollar's value doubled (rising fully 100 
percent, which is unlikely), then we would lose about 7 percent of our manufacturing employment, or 
70,000 jobs. (With about a million manufacturing jobs in the state, each percent lost represents 10,000 
jobs.) This would certainly be a problem, but it cannot be the main problem, for Ohio has lost 400,000 
manufacturing jobs in the last 25 years. Evidently the foreign exchange value of the dollar is only one 
factor among many that have contributed to the secular decline in manufacturing employment. 
25 As usual, a number of caveats apply. Ideally one would hope to control better for 
equilibrium versus disequilibrium changes in the time series, obtain better stationarity, include 
more explanatory variables, and use a multi-equation model. 
119 
7
Haight and Carroll: DOLLAR APPRECIATION AND LAYOFFS IN OHIO MANUFACTURING: TIME SERIE
Published by Carroll Collected,
Journal of Economics and Politics, Vol. 16, No. I, 2003 - 2004. 
REFERENCES 
Bartolini, L. (1995). "Purchasing Power Parity Measures of Competitiveness." Finance and 
Development, 32(3), 46-49. 
Brenner, R. (2002). The Boom and the Bubble: The US in the World Economy. New York: Verso Press. 
Cassell, M. and A. Hanauer (1999-2000). "A Study of Wages and Inequality in Ohio from 1979 to 
1997." Journal of Economics and Politics, 13(1 ). 
Goldberg, P. K. and M. M. Knetter (1997). "Goods Prices and Exchange Rates: What Have We 
Learned?" Journal of Economic Literature, 35, 1243-72. 
Rennie, H. G. (1997). "Simulating the Economic Effects of a Plant Expansion or Contraction: Jeep and 
the Toledo Area Economy." The Ohio Journal of Economics and Politics, 11(1). 
Sachs, J. D. and Shatz, H.J. (1994). "Trade and Jobs in US manufacturing." Brookings Papers on 
Economic Activity, I, 1-69. 
120 
8
Submission to The Journal of Economics and Politics
https://collected.jcu.edu/jep
DOI: 10.59604/1046-2309.1162


English

DOLLAR APPRECIATION AND LAYOFFS IN OHIO MANUFACTURING: TIME SERIES EVIDENCE

Wiggam, Scott S.

CONGRESSIONAL PERSPECTIVES ON THE INFLUENCE OF BUSINESSES THROUGH GOVERNMENT RELATIONS: CONGRESSIONAL STAFF MEMBERS FROM THE OHIO DELEGATION COMMENT ON EFFECTIVE STRATEGIES

Benedict, Mary Ellen

Wilder, Lisa

Krishnan, Sumangali

DO CANDIDATES NEED TO LOSE A STATEWIDE ELECTION

Front Matter

Reilly, Shauna

The Journal of Economics and 
Politics 
Manuscript 1184 
Partisan Identification of Naturalized Americans 
Shauna Reilly 
Follow this and additional works at: https://collected.jcu.edu/jep 
1
Reilly: Partisan Identification of Naturalized Americans
Published by Carroll Collected,
2
Submission to The Journal of Economics and Politics
https://collected.jcu.edu/jep
DOI: 10.59604/1046-2309.1184
3
Reilly: Partisan Identification of Naturalized Americans
Published by Carroll Collected,
4
Submission to The Journal of Economics and Politics
https://collected.jcu.edu/jep
DOI: 10.59604/1046-2309.1184
5
Reilly: Partisan Identification of Naturalized Americans
Published by Carroll Collected,
6
Submission to The Journal of Economics and Politics
https://collected.jcu.edu/jep
DOI: 10.59604/1046-2309.1184
7
Reilly: Partisan Identification of Naturalized Americans
Published by Carroll Collected,
8
Submission to The Journal of Economics and Politics
https://collected.jcu.edu/jep
DOI: 10.59604/1046-2309.1184
9
Reilly: Partisan Identification of Naturalized Americans
Published by Carroll Collected,
10
Submission to The Journal of Economics and Politics
https://collected.jcu.edu/jep
DOI: 10.59604/1046-2309.1184
11
Reilly: Partisan Identification of Naturalized Americans
Published by Carroll Collected,
12
Submission to The Journal of Economics and Politics
https://collected.jcu.edu/jep
DOI: 10.59604/1046-2309.1184


Partisan Identification of Naturalized Americans

Wheeler, Darren

The Value of State Political Science Organizations

Prusak, Bernard

Socially Responsible Investing - the Catholic Way

Berg, Matthew P.

Günter Bischof and Peter Ruggenthaler, Österreich und der kalte Krieg: Ein Balanceakt zwischen Ost und West (Graz/Wien: Leykam, 2022)

Cicirelli, Marianne

DORIAN GRAY MEETS THE DIGITAL AGE

Ellerbruch, Julia

Accounting Firms’ Concerns with the SEC\u27s Proposal to Report Scope 3  Greenhouse Gas Emissions