A Specific Factor Model of FTAA and North Carolina Textile and Apparel Industries

Mostafa Malki

University of Texas - Brownsville

Henry Thompson

Auburn University

Osei-Agyeman Yeboah

North Carolina A&T State University

Textile and apparel industries in the US face import competition that promises to increase under the Free Trade Area of the Americas. The present paper utilizes a specific factors model of production and trade to predict the potential impact of FTAA on the textile and apparel industries in North Carolina. Income is redistributed across six labor skill groups in North Carolina, and returns to capital in textiles and apparel fall as does output. In spite of falling prices for textiles and apparels, the model predicts higher wages based on rising prices of other products.

The Free Trade Area of the Americas (FTAA) is a pending free trade agreement to include all of the Americas patterned after the North America Free Trade Agreement (NAFTA). Smaller US apparel producers have fought a losing battle of import competition in NAFTA and larger US textile manufacturers have expanded into Mexico given the rules of origin and yarn forward requirements. US textile export revenue from Mexico more than tripled between 1994 and 1999 with apparels then re-exported to the US. More recently, US textile exports to Canada and Mexico are falling off and apparel imports from China and the Caribbean Basin are increasing.

The FTAA will change patterns of trade in various textile and apparel products. South America has major cotton growing regions, relatively cheap labor, and growing demand. FTAA will raise average incomes throughout the hemisphere but import competing industries will lose as resources shift toward export production. Income will be redistributed between labor groups, capital, and other inputs. Import competition will intensify for various US textile and apparel products but export demand will expand for other products.

The textile and apparel industries are key components of the North Carolina economy, the largest textile state in the US and the fourth largest apparel state with 33% and 8% of US employment in those industries. In 1995 North Carolina had over 2,000 plants employing about 250,000 but within a decade the number of plants declined by almost 40% and employment by 61%.

The present paper projects adjustments in North Carolina to FTAA in a general equilibrium model of production of the state economy that focuses on the textile and apparel industries. The model assumes six labor skill groups along with energy and industry specific capital inputs, generating general equilibrium comparative static adjustments in outputs and factor prices to projected product price changes due to FTAA. Outputs also adjust to changing output prices in the general equilibrium. The specific factors model is typically applied to national economies but the present study applies it to a state economy.

The effects of changing prices of traded products on factor prices depend on factor intensity and substitution. The paper examines sensitivity of comparative static results to various degrees of constant elasticity substitution.

1. A Review of the Specific Factors Model

The comparative static model of competitive production and trade developed for instance by Jones and Scheinkman (1977) and Chang (1979) is part of the foundation of trade theory. The model assumes constant returns, full employment, and competitive pricing. Each industry in the state is assumed to have its own specific capital input in the present specific factors model with perfect mobility of labor and energy input across industries.

Full employment is stated Ax= v where A is the matrix of cost minimizing unit inputs, x is the output vector, and v is the input vector. Competitive pricing in each industry is stated A^Tw = p where w is a vector of factor prices and p a vector of product prices. The state economy is assumed to be a price taker in product markets.

Full employment leads to the first equation in (1) and competitive pricing to the second. Aggregate substitution terms s_ih summarize substitution of factor i when the price of factor h changes. Letting ¢ represent percentage change, the model in elasticity form is

s l w¢ v¢

= (1)

q^T 0 x¢ p¢

where l is the matrix of industry shares, q the matrix of factor shares s the matrix of substitution elasticities. Factor prices w and outputs x are endogenous in the model while factor endowments v and prices p are exogenous. In the small open state economy, exogenous prices change due to import competition and export expansion under FTAA.

2. Factor and Industry Shares in North Carolina

Factor shares q are the portions of industry revenue paid to productive factors and industry shares l the portions of factors employed by industry, derived as in Thompson (1996). The present model is based on labor data for five skilled groups (managers, professionals, service, agriculture, production) across manufacturing industries, services, and agriculture from the BLS (Bureau of Labor Statistics, 2000). Energy spending by industry is from the US Department of Energy (1998). Value added for agriculture is from the ERS/USDA (1999) and for manufacturing from the US Census of Manufacturers (1997). Value added in services is the residual of state output. Capital receives the residual of value added after the labor and energy bills. Due to the lack of data for energy input in the service sector, the model uses 2% which is the smallest energy factor share in manufacturing.

The value of factor i input in industry j is w_ij º w_iv_ij where w_i is the price of factor i and v_ij the quantity of factor i used in industry j. The share of factor i in industry j is then q_ij º w_ij/y_j where y_j is the value added of industry j. Factor index i runs across industry capital k, energy e, and the five skilled types of labor. The six industries are

G agriculture S services

T textile mills P textile products

A apparel manufacturing M other manufacturing

and the inputs are

L_m managers L_p professional labor

L_s service labor L_c clerical labor

L_a agricultural labor L_w production labor

K_j industry specific capital E energy.

Table 1 is the total factor payment matrix. Agricultural workers are industry specific, service workers almost so, and about ¾ of production workers are in other manufacturing M. North Carolina manufacturing accounts for about ½ of value added, high relative to the US. Textile mills are just over 11% as large as the rest of manufacturing, apparel about half as large as textile mills, and textile products less than half as large as apparel. Textile and apparel industries sum to about 9% of state value added.

Table 1. Factor payments ($mil)

	G	S	M	T	P	A	income
Lm	43	20,264	2,805	248	37	148	23,544
Lp	89	32,995	2,187	55	12	67	35,406
Ls	84	28,974	0	0	0	0	29,058
Lc	0	13,972	136	110	15	27	14,261
La	200	0	0	0	0	0	200
Lw	467	0	11,552	2,346	425	1,103	15,892
ΣjKj	3,080	23,201	110,610	14,434	2,245	5,843	159,414
E	1,094	28,754	6930	171	35	47	37,032
value added	5,057	148,162	134,220	17,365	2,769	7,235	314,807

Summing a column in Table 1 gives total industry revenue. The total revenue of services in Table 1 is $148 billion making the capital share $23.2/$148 = 15.7%. Table 2 presents the derived factor shares. Capital has the largest factor share except in services, and the largest factor shares in the textile and apparel industries are for production workers.

Table 2. Factor shares q_ij

	G	S	M	T	P	A
Lm	.008	.137	.021	.014	.013	.020
Lp	.018	.223	.016	.003	.004	.009
Ls	.017	.196	0	0	0	0
Lc	0	.094	.001	.006	.005	.004
La	.040	0	0	0	0	0
Lw	.092	0	.086	.135	.154	.152
Kj	.609	.157	.824	.831	.811	.808
E	.216	.194	.052	.010	.013	.006

Industry shares in Table 3 show the distribution of inputs across industries. The sum across rows in Table 1 gives total factor income. Perfect labor mobility within the state implies the wage of each type of labor is the same across industries, and industry shares can be derived. For instance, total income of service workers is $29 billion implying $28.9/$29 = 99.7% of service workers are in the service sector. Professionals and managers also typically work in services. Very large shares of service workers and clerks are in the large service sector, and production workers in manufacturing. Capital industry shares equal one since capital is industry specific.

Table 3. Industry shares l_ij

	G	S	M	T	P	A
Lm	.002	.861	.119	.011	.002	.006
Lp	.003	.932	.062	.002	0	.002
Ls	.003	.997	0	0	0	0
Lc	0	.980	.010	.008	.001	.002
La	1.00	0	0	0	0	0
Lw	.029	0	.727	.148	.027	.069
E	.030	.776	.194	.187	.005	.001

The assumption of perfect mobility of each type of labor is based on the notion that clerks, for instance, could move between industries finding in response to wage differentials. Labor mobility is certainly not perfect. Changing industries may involve moving and workers must consider local amenities in making their job location decision. Assuming a fixed endowment of each type of labor does not allow workers to switch between classifications but in reality there would be at least some mobility of workers across labor types. These issues go well beyond the present competitive model but might not greatly influence the comparative static analysis and conclusions of the simulations.

Textile and apparel industries employ about a quarter of production workers, and textile mills alone employ about 15%. Apparel manufacturing employs about half as many production workers as textile mills, and textile products half of that. The intensity and high employment of production workers in textiles suggests there will be a large effect of FTAA on the demand for production workers.

3. Comparative Static Elasticities in the Applied Specific Factors Model

Substitution elasticities summarize adjustment in cost minimizing inputs when factor prices change as summarized by Jones (1965) and Takayama (1982). Following Allen (1938) the cross price elasticity S_ih^j between the input of factor i and the payment to factor h in industry j is derived from the Allen elasticity E_ih^j as S_ih^j = q_hjE_ih^j. Linear homogeneity implies å_kE_ik^j = 0 and the own price elasticities E_ii^j are derived the negative sum of the cross price elasticities. Cobb-Douglas production implies unit Allen elasticities E_ih^j implying substitution elasticities equal factor shares, S_ih^j = q_hj. Constant elasticity of substitution (CES) production implies the Allen partial elasticity has some positive value. The present simulations apply a range of CES substitution for sensitivity.

Substitution elasticities are the weighted average of industry cross price elasticities, s_ih º å_jl_ijE_ih^j = å_jl_ijq_hjS_ih^j. Factor shares and industry shares are sufficient to derive CES substitution elasticities. Table 4 reports Cobb-Douglas substitution elasticities, and CES scales these elasticities accordingly. For instance, if CES = ½ the substitution elasticities would be half as large. Substitution elasticities in the applied production literature range from ½ to 1. Notation in Table 4 includes L_i for labor inputs and K for industry capital. Energy input is E and e its price. Industry capital returns are r_j.

Table 4. CES substitution elasticities s_ik

	wm	wp	ws	wc	wa	ww	e	rG	rS	rM	rT	rP	rA
Lm	-1.46	.194	.168	.081	0	.013	.249	.001	.726	.021	.002	0	.001
Lp	.129	-1.41	.182	.088	0	.006	.201	.001	.786	.011	0	0	0
Ls	.136	.222	-1.45	.094	0	0	.158	.001	.841	0	0	0	0
Lc	.134	.218	.192	-1.55	0	.002	.170	0	.826	.002	.001	0	0
La	.008	.018	.017	0	-1.14	.092	.609	.391	0	0	0	0	0
Lw	.019	.014	0	.002	.001	-1.04	.817	.011	0	.128	.025	.005	.013
E	.110	.177	.152	.073	.001	.020	-1.23	.012	.655	.033	.001	0	0
KG	.008	.018	.017	0	.040	.092	.216	-.391	0	0	0	0	0
KS	.137	.223	.196	.094	0	0	.194	0	-.843	0	0	0	0
KM	.021	.016	0	.001	0	.086	.052	0	0	-.176	0	0	0
KT	.014	.003	0	.006	0	.135	.010	0	0	0	-.169	0	0
KP	.013	.004	0	.005	0	.154	.013	0	0	0	0	-.189	0
KA	.020	.009	0	.004	0	.152	.006	0	0	0	0	0	-.192

The largest own substitution occurs for clerk wages with every 1% increase leading to over a 1.5% decrease in their input. The smallest own substitution is for capital in textile mills where every 1% increase decreases input by 0.17%. Own elasticities are inelastic for capital but elastic for labor. There is weak cross price substitution.

Table 5 reports the elasticities of factor prices with respect to product prices. These comparative static elasticities are found by inverting the system matrix in (8). Some factors benefit but others lose, and the effects are uneven.

Table 5. Price elasticities of factor prices

	pG	pS	pM	pT	pP	pA
wm	-.002	.983	.016	.001	0	.001
wp	-.002	1.01	-.001	-.001	0	0
ws	-.003	1.02	-.019	-.001	0	-.001
wc	-.005	1.02	-.016	0	0	0
wa	1.05	-.012	-.024	-.005	-.001	-.003
ww	.038	.587	.282	.052	.011	.029
e	.012	.952	.031	.002	.001	.001
rG	1.56	-.497	-.052	-.008	-.002	-.005
rS	-.004	1.03	-.018	-.001	0	-.001
rM	-.005	-.167	1.18	-.006	-.001	-.003
rT	-.006	-.135	-.046	1.19	-.002	-.005
rP	-.007	-.155	-.054	-.010	1.23	-.005
rA	-.007	-.160	-.054	-.010	-.002	1.23

Every 1% increase in the price of agriculture would raise agricultural wages by 1.05%, production wages by 0.38%, and the return to capital in agriculture by 1.56%. Higher agricultural prices increase agricultural output and attract labor from other industries raising its capital productivity. Every 1% increase in the price of other manufactures would raise production wages by 0.28%, wages of managers by 0.02%, and capital returns by 1.18%.

Wages depend very little on prices of textiles and apparel, but heavily on the price of services. Only capital owners in the industry have much at stake with changing textile and apparel prices. Even production workers would suffer less than half of 1% decline in wages with a 10% price decline for textiles and apparel. The reasons for the small effect are increased outputs in other industries and labor mobility across industries.

If labor were industry specific the effect of falling prices would be lower wages with effects on the scale of capital returns. Given labor contracts with fixed wages, the effect would switch to increased unemployment. Certainly these issues are locally important but the effects would be transitory. The present assumption of labor mobility captures inevitable competitive forces.

Thompson and Toledo (2001) show the comparative static effect of price changes on factor prices are the same for any CES production function. The degree of substitution, constant along isoquants, does not affect these general equilibrium price elasticities in competitive models of production. The comparative static elasticities in Table 5 then apply with any degree of CES substitution.

Table 6 reports price elasticities of outputs along the production frontier. A higher price raises its output drawing labor and energy away from other industries. The largest own output effect occurs in agriculture where a 10% price increase raises output 5.64%. Every 1% price increase in textile and apparel raises its output about 0.2%. The smallest own effect is for services because there are fewer resources to attract from the rest of the North Carolina economy.

Table 6. Price elasticities of outputs

	pG	pS	pM	pT	pP	pA
xG	.564	-.497	-.052	-.008	-.002	-.005
xS	-.004	.025	-.018	-.001	0	-.001
xM	-.005	-.167	.182	-.006	-.001	-.003
xT	-.006	-.135	-.046	.194	-.002	-.005
xP	-.007	-.155	-.054	-.010	.231	-.005
xA	-.007	-.160	-.054	-.010	-.002	.233

4. Adjustments to FTAA

NAFTA has had a negative impact on US textile and apparel industries. Predictions across industries varied according to factor intensity with labor intensive industries projected to decline as in Weintraub, Rubio, and Jones (1991), Marchant and Rupel (1993), Boyd, Krutilla, and Kinney (1993), Hansen (1994), Thompson (1996), ERS/USDA (1998a), and Wall (2000).

The present assumptions regarding FTAA price changes are that the price of textiles and apparel will fall along with the average price of agricultural products. The price of other manufactures is assumed to stay constant while the price of services rises. To derive the endogenous vector of factor price adjustments, multiply a vector of predicted price changes by the matrix of factor price elasticities in Table 5.

Table 7 assumes a range of FTAA price changes in the FTAA Prices columns for some perspective on potential adjustments. While the price changes are arbitrary, results scale to the level of price changes and the factor price adjustments in the second set of columns are identical for any degree of CES.

Table 7. Factor price and output adjustments to FTAA Prices

	% FTAA Prices		% Factor prices				% Outputs			% Long run outputs
pG	0 -2 -5	rG	.15	-3.2	-8.0	xG	.15	-1.2	-3.0	.15	-3.2	-8.0
S	0 1 2	rS	.03	1.1	2.3	xS	.03	0.1	0.2	.03	1.1	2.3
pM	-1 -2 -5	rM	-1.1	-2.3	-5.9	xM	-.12	-.29	-.86	-1.1	-2.3	-5.9
pT	-5 -20 -30	rT	-5.8	-24	-35	xT	-.86	-3.7	-5.5	-5.8	-24.0	-35
pP	-5 -20 -30	rP	-6.0	-24	-36	xP	-1.0	-4.3	-6.4	-6.0	-24	-36
pA	-10 -30 -50	rA	-12	-37	-61	xA	-2.2	-6.8	-11	-12	-36	-61
		wm	-.03	.91	1.8
		wp	.01	1.0	2.1
		ws	.03	1.1	2.2
		wg	.02	1.1	2.2
		wa	.08	-1.9	-4.8
		ww	-.89	-2.2	-3.8
		e	-.06	.78	1.6

Wage adjustments are generally very small due to the mobility of labor across industries, and are linked to factor and industry shares. Only when prices change in the range of 20% and upward are there noticeable wage effects. Capital returns, however, adjust to a larger extent than their price changes due to the magnification effect. Adjustments in the price of energy e are negligible.

Output adjustments in Table 7 are found by multiplying output elasticities in Table 6 by the vectors of projected price changes. Output adjustments are much less than price changes in the short. Output adjustments scale with CES substitution and would be half as large with CES = ½ and estimates of substitution in the literature generally fall in the range of ½ to 1. Changes in outputs also scale to their price changes.

It is reasonable to assume the price of textile mill products would rise but apparel prices would fall under FTAA. Output and capital returns generally follow price changes. The present model suggests output adjustments would be smaller but capital return adjustments larger than industrial price changes.

A decreased capital return will lower investment and generate larger long run output adjustments. To examine the long run potential output adjustment, assume the elasticity of capital with respect to its return equals one as is approximately the case under the present model. In the production adjustment, the percentage change in industry output would then be about equal to the percentage change in its capital stock. These capital stock changes magnify output adjustments in the long run. The last column of Table 7 shows these long run output changes and the declines in textile and apparel outputs are sizeable.

Factor price changes are proportional to the vector of price changes. For instance, a doubling of price changes in a particular vector leads to factor price and output adjustments twice as large as those in Table 7.

Results are insensitive to the assumption of specific capital. The various types of labor are nearly specific to their particular industries: managers, professionals, and service workers in services, agricultural workers in agriculture, and production labor in other manufactures. Price changes have similar effects on income distribution regardless of whether capital is industry specific.

5. Conclusion

The present applied specific factors model projects the range of income redistribution and output changes in North Carolina as it adjusts to FTAA. The North Carolina textile and apparel industry will suffer import competition under FTAA but there will be rising prices and export opportunities for production intensive in skilled labor.

Output and wage adjustments will not be overwhelming under reasonable price scenarios but adjustments in capital returns are magnified effects of price changes. Wage adjustments are small due to the assumption of labor mobility across industries, and would be magnified effects of price changes if labor were immobile between industries. Labor could retrain in response to changing wages for particular skills and labor mobility between skill groups would diminish the wage impacts.

Output adjustments will be negligible in services and other manufacturing. Wages of all but agricultural and production workers rise in FTAA while returns to capital fall in the textile and apparel as well as other manufacturing industries. Output adjustments are smaller than their price decreases in the short run but larger in the long run due to declining investment, and textile and apparel output will substantially fall in the long run.

References

Allen, R.G.D. (1938) Mathematical Analysis for Economists. MacMillan.

Boyd, R.G., K. Krutilla, and N.C. Kinney (1993) “The Impact of Tariff Liberalization between the United States and Mexico: An Empirical Analysis,” Applied Economics 22, 81-89.

Bureau of Labor Statistics (2000) web page.

Chang, Winston (1979) “Some Theorems of Trade and General Equilibrium with Many Goods and Factors,” Econometrica 47, 709-726.

ERS/USDA (1999) Farm Income and Costs: Farm Income Forecasts.

ERS/USDA (1998a) “Free Trade in the Americas,” International Agriculture and Trade Reports, Situation and Outlook Series, WRS-98-1; 54.

ERS/USDA (1998b) “Free Trade in the Americas: Potential Advantages for U.S. Agriculture,” Agricultural Outlook, April: 11-15.

Hansen, N. (1994) “NAFTA: Implications for U.S. Regions,” The Survey of Regional Literature, 4-22.

Jones, Ron (1965) “The Structure of Simple General Equilibrium Models,” Journal of Political Economy, 557-572.

Jones, Ronald and José Scheinkman (1977) “The Relevance of the Two-Sector Production Model in Trade Theory,” The Journal of Political Economics 15, 65-99.

Marchant, M. A. and F.J. Rupel (1993) “The Impacts of Lesser Developed Countries on Southern Region Agricultural Exports,” Agricultural and Resource Economic Review 22, 71-87.

Takayama, Akira (1982) “On the Theorems of General Competitive Equilibrium of Production and Trade – A Survey of Some Recent Developments in the Theory of International Trade,” Kieo Economic Studies, 1-37.

Thompson, Henry (1996) “NAFTA and Industrial Adjustment: Specific Factor Model of Production in Alabama,” Growth and Change, 3-28.

Thompson, Henry and Hugo Toledo (2001) “Bolivia and South American Free Trade,” International Trade Journal, 113-126.

US Census of Manufacturers (1997) Manufacturing by Subsector.

US Department of Energy (1998) Expenditures for Purchased Energy Sources, 1998; National Data and Regional Totals.

Wall, Howard (2000) “NAFTA and the Geography of North American Trade,” Federal Reserve Bank of St Louis Working Paper, www.stls.frb.org/docs/research/wp/2000-017.pdf.

Weintraub, S, .L. Rubio, and A. Jones (1991) US - Mexican Industrial Integration: The Road to Free Trade, Boulder: Westview Press.