Trans-Pacific Partnership Agreement Appendix G

Introduction

The discussion that follows focuses on the quantitative analysis in this report $—$ the computable general equilibrium (CGE) analysis presented in chapter 2 and the industry estimates presented in chapters 3, 4, and 5. This appendix details the procedures used to adapt the standard Global Trade Analysis Project (GTAP) model to enable the Commission to assess the likely effects of TPP. The basic features of the GTAP model are introduced, along with a discussion on adjustments made to the standard database, the development of the baseline, and the various analyses incorporating the different TPP provisions quantified in the model, including the liberalizations in tariffs, certain nontariff measures, and investment restrictions.

The GTAP Model

The GTAP project has two main components. One is a documented global database on international trade, economy-wide inter-industry relationships, and national income accounts (the GTAP database). The other is a standard modeling framework to organize and analyze the data (the GTAP model). The modeling framework allows comparisons of the global economy in two environments: one in which the base values of policy instruments such as tariffs, tariff-rate quotas (TRQs), or export restrictions are unchanged, and one in which these measures are changed, or “shocked,” to reflect the policies that are being studied. A change in policy makes itself felt throughout the economies depicted in the model. We begin with the latest release of the GTAP database, version 9, and start with the GTAP model and assumptions as discussed in the USITC’s analysis of the U.S. $-$ Korea free trade agreement.[1]

Results from the GTAP model are based on established global trade patterns. This means that the model is unable to estimate changes in trade in commodities that historically have not been traded. That is to say, if a particular commodity is not traded between two economies, no model simulation will bring about such a trade flow under any circumstance. Furthermore, patterns of trade may exist for such reasons as the distance between countries, the presence or absence of transport infrastructure, or cultural preferences, which are all imperfectly captured by the model. The GTAP model does not directly account for historical or cultural factors as determinants of trade patterns. The model assumes that these factors are unaffected by the trade policy change.

In the GTAP model, domestic products and imports are consumed by firms, governments, and households. Product markets are assumed to be perfectly competitive (implying zero economic profit for the firm).[2] In the model, imports are imperfect substitutes for domestic products (i.e., consumers are aware of the source of the products and may distinguish between them based on the foreign or domestic origin), and sectoral production is determined by global demand and supply.

The Dynamic GTAP Model

The CGE simulation model used in this report can also be seen as consisting of two parts. The first part is the standard static CGE model, as discussed in the previous section. The model simulates changes, assuming that the economy-wide supply of labor responds to changes in the real wage rate while the supplies of all other primary factors are fixed. The static model by design does not produce information about the speed with which changes occur or about what happens to various dimensions of the economies in the meantime. Rather, the simulation finds the new equilibrium of prices and quantities within the model that result in response to the change in policy.

The second part of the CGE model provides for dynamic linkages and simulates changes over time. To simulate changes in the structure of the U.S. economy over time, the simulation framework incorporates physical capital accumulation for the economy as a whole. Capital accumulates each period as new investment, less depreciation on existing assets, adds to the capital stock.

The level of new capital goods or investment is determined by the static model. The Commission uses a “baseline” which describes the expected evolution of the world economy in the absence of the TPP Agreement. The baseline runs from 2017 to 2047 in five-year steps and incorporates projections for labor availability, growth rates for population and gross domestic product (GDP), and trade policy changes that would take place in the absence of TPP $—$ for example, the tariff changes provided in the recently signed Japan-Australia Economic Partnership Agreement. Population and labor availability are exogenous variables in the CGE model. Thus these variables are shocked in every period, according to the projections. GDP, however, is normally an endogenous variable in the CGE model. To target GDP, the closure of the model is changed, with GDP growth made exogenous, and an economy-wide technology parameter allowed to adjust as needed. The baseline incorporates projections from the International Monetary Fund (IMF) and Organisation for Economic Co-operation and Development (OECD) for labor, population, and GDP growth rates. Table G.1 shows forecast growth in real GDP, labor force, and population adopted in the baseline; table G.2 shows selected results from incorporating these forecasts.

Table G.1: Inputs to projected U.S. baseline: Five-year cumulative growth rates for U.S. real GDP, labor force, and population, percent

Time period	Real GDP	Labor force	Population
2017 $-$ 22	13.64	2.37	3.55
2022 $-$ 27	12.60	2.44	3.29
2027 $-$ 32	11.91	2.65	2.88
2032 $-$ 37	10.64	2.96	2.48
2037 $-$ 42	9.45	3.01	2.17
2042 $-$ 47	8.56	2.78	1.96

Source: USITC estimates.

Table G.2: Selected indicators from projected U.S. baseline: Five-year cumulative growth rates for U.S. capital stock, real private consumption, real exports, and real imports, percent

Time period	End of period capital stock	Real private consumption	Real exports of goods and services	Real imports of goods and services
2017-22	8.37	9.39	10.35	12.56
2022-27	13.82	11.48	9.12	11.05
2027-32	15.04	10.89	9.21	10.78
2032-37	15.41	10.00	8.54	10.03
2037-42	15.02	8.60	7.94	9.95
2042-47	14.80	7.19	7.45	10.27

Source: USITC estimates.

The simulation of the TPP Agreement then generates a “policy” line. The policy simulations include the TPP-related policy changes and several variables used in the baseline simulations, including population and labor growth and the economy-wide technology parameter. For a particular variable, e.g., total U.S. exports, the distance between the “policy” line and the “baseline” is the effect of the TPP Agreement. The TPP shocks that we simulate in this report start in the year 2017. That is, we assume 2017 as the year of entry into force and the first year that the policy line deviates from the baseline.

Updating and Modifying the GTAP Database

As noted earlier, for the purpose of the TPP analysis, the Commission has updated and modified the standard GTAP database to reflect current U.S. and global economic conditions, and to project future U.S. and global economic conditions both under TPP and in the absence of TPP.

The current standard version of the GTAP database (version 9) contains 140 regions and 57 sectors. The standard GTAP data are based on the year 2011 $—$ that is, figures for trade flows, trade barriers, and other data refer to the world in that year.

In addition to the data on bilateral trade found in each of the sectors in the model, data are incorporated on the domestic production and use of output in each sector (including its use in the production of other commodities and services); the supply and use of land, labor, and capital; population; and GDP. The database also contains information on tariffs, some nontariff barriers, and other taxes. An additional component of the data is a set of parameters which, in the context of the model’s equations, determine economic behavior. These are principally a set of elasticity values that determine, among other things, the extent to which imports and domestically produced goods are substitutes for one another.

TPP Model Regions and Sectors

The Commission’s analysis focuses on U.S. trade with TPP members and other important trading partners to the United States. Table G.3 shows the countries and regions specified in the model. They include the 12 TPP parties, China, the European Union (EU), Hong Kong, Indonesia, South Korea, Thailand, and the rest of the world as a region.

Table G.3: Model regions

TPP parties		Other countries/regions
United States	Australia	China
Canada	New Zealand	EU
Mexico	Malaysia	Hong Kong
Chile	Singapore	Indonesia
Peru	Vietnam	South Korea
Japan	Brunei	Thailand
		Rest of the world

Source: Compiled by USITC staff.

Also, the GTAP database’s 57 sector aggregation was modified, making it possible to focus on particular industries of interest. In total, 56 industry sectors are specified in the model, including both goods and services. Table G.4 lists all the model sectors.

Table G.4: Model goods and services sectors

Model sectors
Rice	Poultry meat products	Auto parts and trailers
Wheat	Soybean oil	Other transportation equipment
Other grains	Soybean meal	Electronic equipment
Corn grain	Dairy products	Instruments and medical devices
Fresh fruit, vegetables, and nuts	Sugar, sweeteners, and SCPs	Toys, sporting goods, and other manufacturers
Soybeans	Processed foods	Electricity
Other oil seeds	Chemicals	Gas manufacture, distribution
All other agriculture	Beverages and tobacco products	Water
Cattle, sheep, goats, and horses	Textiles	Construction
Hides and skins	Wearing apparel	Wholesale and retail trade
Forestry	Leather products	Transportation, logistics, travel and tourism
Seafood	Footwear	Communications
Coal	Wood products	Financial services n.e.c.
Oil	Paper products, publishing	Insurance
Gas	Petroleum, coal products	Business services n.e.c.
Minerals and minerals products n.e.c.	Machinery and equipment	Recreational and other services
Beef meat	Metal products n.e.c.	Public administration, defense, education, health
Other meats	Titanium downstream products	Dwellings
Pork meat products	Passenger vehicles

Source: Compiled by USITC staff.

Note: N.e.c. = not elsewhere classified; SCPs = sugar-containing products.

Updating the Database

For the purpose of the present study, a number of updates have been made to benchmark data on trade flows and GDP growth. The model is then projected to 2017 using estimates of regional and global GDP growth. Data are drawn from the U.S. Department of Commerce (U.S. imports and exports), the IMF’s World Economic Outlook (GDP projections). Observed GDP growth rates for all the regions of the model are targeted using these data, as are population growth rates. Trade flows within the model are adjusted to reflect key observable trade in the real world. The strategy employed is to match disaggregated trade flows that are critical to the results of the policy simulation. Once the database is updated to align with key observed 2014 data, the 2014 database is then projected forward to 2017. This is accomplished by incorporating real GDP and population growth projections from the sources listed above.

Key Assumptions

The Commission’s simulation results depend not only on the GTAP model and parameters, but also on a number of assumptions made to align the baseline and policy simulations with the Commission’s interpretation of the Agreement and with economic forecasts necessary to establish the baseline. The Commission’s simulations do not currently incorporate any adjustment costs. This assumption means that the sectoral allocation of labor may change without any additional costs to workers or firms.

To align the model with anticipated demographic changes in the United States and elsewhere in the world, population growth, labor force, and labor force participation are based on annual data and forecasts through 2050 published by the International Labour Organization (ILO). Forecast growth in real GDP were obtained from the IMF’s World Economic Outlook databases and the OECD’s long-term baseline projections.

In addition to labor force projections, participation rates were modeled to respond to changes in real wages. As real wages (wages adjusted for the price index for private consumption) rise, the labor supply expands accordingly. The real labor supply elasticities for both skilled and unskilled labor are 0.4 for the United States (published by the CBO) and other developed economies, while 0.44 is used for other economies (based on a review of the literature).

The ratio of the U.S trade deficit (that is, exports less imports) relative to GDP was kept fixed within the model by allowing changes in the U.S. savings rate to fluctuate.[3]

In addition to the specification and modeling of provisions regarding foreign direct investment and NTMs affecting traded services, discussed later in this appendix, the following assumptions were also made. Trade responses for U.S. exports of dairy products to Canada and Japan follow expansions in the corresponding quota levels, implying that the United States will take full advantage of future quota expansions. U.S. imports of cheese from New Zealand were modeled not to expand after U.S. tariff changes because the Commission has determined that U.S. and New Zealand cheese are not comparable products. In sugar, U.S. exports and imports follow expansions in the corresponding quota levels.

U.S. beef meat exports to Japan were modeled to reflect the preference of Japanese consumers for Japanese beef meat. U.S. exports of meat products to Malaysia were modeled not to expand because of limited available expansion capacity for Malaysian-approved Halal meat plants in the United States. U.S. poultry product exports to and imports from Canada were modeled not to respond significantly to Canadian tariff changes because U.S. exports to Canada are essentially duty free (due to duty drawbacks) and these products are reexported to the United States after being processed in Canada.

U.S. trade responses to tariff changes in textiles, apparel, leather products, and footwear were modeled to reflect existing supply chain relationships and capacity constraints.

The existing regime of duty drawback in Vietnam generally reduces the effect of duty reductions and removals. Vietnamese trade was modeled to reflect the influence of TPP rules of origin. For Vietnam to take advantage of reduced tariffs on its products, Vietnam was modeled to prefer trading more with TPP economies and less with non-TPP economies.

For the sectors including instruments and medical devices; other transportation equipment; and other machinery and equipment, U.S. exports to non-TPP countries were modeled to reflect modest changes due to the competitive advantages of U.S. exporters of those products.[4]

Alternative Model Assumptions

Certain assumptions and policy changes to the model discussed above were introduced based on industry expertise. These inputs include the degree of substitution between domestic and foreign varieties of certain goods and the expected restrictiveness of select TRQs, among other factors. Economy-wide effects excluding this information are shown in table G.5 below.

Table G.5: Aggregate effects of TPP liberalization under alternative model assumptions

Measure	2032		2047
	Billion $	Percent		Billion $	Percent
Real income	60.5	0.25		88.3	0.30
Real GDP	44.1	0.16		68.8	0.18
Employment, FTE thousand	128.8	0.07		176.0	0.09

Total exports	27.8	1.0
Agriculture and food	11.0	4.1
Manufacturing, natural resources, and energy	12.0	0.7
Services	4.9	0.6
Total imports	51.8	1.2
Agriculture and food	5.3	3.6
Manufacturing, natural resources, and energy	39.5	1.1
Services	6.9	1.2

Source: USITC estimates.

Incorporating Market Access Provisions

In order to understand the incremental effects of the market access provisions under TPP (Chapter 2 of the TPP text), two databases of tariffs from 2017 through 2046 were constructed to show the evolution of tariffs absent and including TPP.

Tariffs and TRQs in the Absence of TPP

To calculate TPP’s potential effects on trade, the model was updated with most-favored-nation (MFN) tariff rates[5] from TPP member countries’ 2014 tariff schedules, and with current and future preferential rates given to other TPP members under pre-TPP free trade agreements (FTAs), such as the North American Free Trade Agreement, or NAFTA. These data were gathered from national authorities (e.g., the Commission, Japan Customs, Canada Border Services Agency, etc.) at the national tariff line level, and were processed by Market Analysis and Research, International Trade Centre UNCTAD/WTO. The MFN rates were assumed to remain unchanged throughout the simulation horizon. Rates for existing FTAs were extrapolated after their respective full implementation.[6] Where rates were expressed as specific or compound rates, ad valorem equivalents were calculated using the WTO-World Tariff Profiles methodology.[7] For lines subject to TRQs, information about the fill rate[8] was used to determine whether the rate to be charged against imports for that product would take the in-quota rate or the out-of-quota rate.

Data were aggregated in two steps. First, to fully account for existing preferences, if multiple import programs applied to the same tariff line, rates were selected for each tariff line by choosing the lowest rate given to imports under all existing import programs (comparing MFN with existing FTA rates, if any). National tariff lines were then aggregated by simple averages to the Harmonized System (HS) 6-digit level, the level at which international tariff rates are published. Finally, tariff rates were aggregated to the sector levels found in the model using three-year averages of bilateral trade as weights. The trade data came from three different sources: Trade Map, IDB, and Comtrade.[9]

Tariffs and TRQs under TPP

TPP tariff schedules were processed according to the text of the agreement. National tariff schedules in the agreement’s text were first processed to show tariff levels throughout the implementation of the agreement. This information was aggregated by simple averages to the HS 6-digit level; information about TRQs and specific tariffs were provided by Market Analysis and Research, International Trade Centre UNCTAD/WTO.[10] Tariff data under the TPP were replaced with tariff data without the TPP in situations where tariffs under the TPP would be higher than other existing FTA rates.[11] Finally, the data were aggregated in the same manner as the database without TPP tariff rates.

Rules of Origin

The Commission simulations include modeling of provisions regarding rules of origin for Vietnamese exports of textiles and apparel. In particular, the simulations are run under the constraint that Vietnamese exporters may expand their exports under reduced tariffs only if they increase their use of originating intermediate inputs and reduce their use of non-originating intermediate inputs.

Incorporating Services Liberalization

The TPP Agreement contains market access provisions that liberalize cross-border trade in services with TPP partners, and national treatment provisions that enable firms to establish commercial presence in TPP partner markets more easily.

Market-access provisions for services are found in TPP’s Chapter 10, Cross-Border Trade in Services; Chapter 11, Financial Services; and to a limited extent in Chapter 13, Telecommunication Services. National treatment provisions related to services firms established abroad are included in TPP’s Chapter 9, Investment, and in both the Financial Services and the Telecommunications chapters. In addition, provisions on MFN treatment, restrictions on local-presence requirements, and obligations regarding transfers would prevent discrimination against foreign services suppliers. Where TPP partners wish to retain certain nontariff measures in a particular sector, rather than committing to full liberalization in the sector, these are noted as nonconforming measures (NCMs) and are listed in Annexes I $-$ III of the agreement.

This section describes the analysis conducted to assess the impact of the TPP Agreement on cross-border services trade. The effects of TPP on services trade that is provided via commercial presence (mode 3) is considered in our analysis on the effects of the agreement on foreign affiliate sales, described later in this appendix.

Estimated Trade Costs for Cross-border Services Trade

To analyze the effects of a possible liberalization in cross-border services trade under TPP, it is necessary to understand the level of existing barriers by country and services sector. One way to summarize these barriers in a country is to estimate their effects in raising the costs to import such services. These costs can be expressed in AVEs (i.e., as a rate equal to a percentage of a traded service’s value) and are often referred to as “tariff equivalents.” The tariff equivalents used in this analysis are taken from the empirical literature on services barriers.[12] These are derived from a gravity approach for each services sector in the GTAP database.

For each services sector, the estimating equation in this analysis takes the form
$x_{i j} = c + δ y_{j} + α_{i j} {dist}_{i j} + β_{ij} D_{i j} + γ_{i} + γ_{j} + ε_{i j}$

where $x_{ij}$ represents the log of exports from country $i$ to partner $j$ . Trade costs other than regulations between $i$ and $j$ are proxied by ${dist}_{i j}$ , the log of their bilateral distance. The vector $D_{i j}$ contains bilateral trade determinants common in the gravity literature, including common language, engagement in a FTA, etc., controlled by indicator variables.

Exporter and importer fixed effects ( $γ_{i}$ and $γ_{j}$ respectively) are included in the model to account for the usual multilateral resistance terms. Without longitudinal data, measures of output and expenditure collapse in the country fixed effects. To disentangle the importer expenditure from the degree of restrictiveness of trade, exports are normalized by the potential size of the market, and the coefficient for $y_{j}$ is fixed.[13] The last term in the equation, $ε_{i j}$ , represents an error term.

The estimation of the gravity equation above is done using the latest GTAP database released in 2015, which provides data for bilateral trade in services by broad services sectors for the year 2011.[14] This estimation is conducted for the following sectors: construction (cns); communication (cmn); trade (trd); finance (ofi); other services (osg), comprising education, health, defense, and public administration; business (obs); water transport (wtp); air transport (atp); other transport (otp); and insurance (isr).

Derivation of Tariff Equivalents

Tariff equivalents $t_{j}$ are obtained from the estimated gravity model. The relationship used to derive these is:
$\ln {(1 + t_{j})}^{1 - σ} = γ_{j} - γ_{j *}$

which relies on the estimated fixed effect $γ_{j}$ for importer country j, relative to $γ_{j *}$ , the fixed effect for a “benchmark” importing country or the country with the largest fixed effect (e.g., Luxembourg in the estimation for communication services).[15] The specific values of the tariff equivalents would also depend on the elasticity of substitution , which is not estimated in the model, but needs to be assumed. The empirical literature on gravity suggests that this elasticity could range in values from 5 to 10.[16] An intermediate value of 8 is assumed. A higher provides lower AVEs, and vice versa. The relative ranking among the different countries, however, is not sensitive to the assumed value of the elasticity of substitution.

Table G.6: Estimated ad valorem equivalent of trade costs by party and services sector, percent, 2011

	cmn	cns	isr	obs	ofi	osg	trd	otp	atp	wtp
Canada	37.0	49.4	36.6	29.0	43.9	41.3	36.5	20.9	20.9	39.4
Mexico	68.4	85.9	16.7	85.2	79.1	47.7	45.2	32.7	32.7	88.2
Chile	45.2	69.3	42.4	45.9	50.4	45.5	36.1	18.2	18.2	17.7
Peru	48.2	38.8	47.2	46.0	76.4	50.6	59.5	36.8	36.8	64.1
Japan	60.2	23.8	51.4	35.5	61.4	54.1	35.1	24.1	24.1	30.1
Australia	44.8	71.2	53.8	39.2	63.1	45.0	42.2	21.5	21.5	45.7
New Zealand	32.3	32.1	42.4	28.8	49.4	37.6	30.9	14.5	14.5	27.4
Malaysia	20.1	8.3	34.5	18.7	46.7	34.7	33.5	14.4	14.4	27.0
Singapore	12.1	31.3	15.1	7.6	24.2	27.1	8.7	0.0	0.0	7.3
Vietnam	29.0	21.5	37.4	32.5	43.6	36.1	35.9	25.5	25.5	37.8
Brunei	49.2	16.1	56.7	31.0	60.3	21.9	31.4	20.6	20.6	32.8

Source: USITC calculation based on Fontagné, Mitaritonna, and Signoret, “Estimated Tariff Equivalents,” 2016.

TPP Liberalization

Services trade is liberalized in TPP through a number of different avenues. The three primary routes are (1) commitments to reduce or remove specific nontariff measures which had been reserved exceptions (NCMs) in previous trade agreements; (2) adoption of a negative list approach (discussed below) to commitments; and (3) the adoption of broad disciplines on ensuring the ability to transmit data across borders and on prohibiting the introduction of data-localization measures (requirements that data be stored and/or processed only in-country). Other parts of TPP also introduce helpful disciplines for services trade $—$ intellectual property protections, rules about state-owned enterprises, government procurement rules, and commitments to improve regulatory coherence, for example $—$ but the impact of these taken together is judged to be less important than that of the three primary factors listed above.[17] The methodology adopted to incorporate the TPP effects of these three factors is outlined below.

A TPP party may commit to a more liberal services trade regime than it did under the General Agreement on Trade in Services or in prior FTAs. For instance, the party may remove an item from its list of NCMs, or propose an NCM which is not as wide ranging as it was previously. To represent these changes quantitatively, the World Bank’s Services Trade Restrictiveness Index (STRI) for the country-sector-mode combination was rescored to reflect the new policy setting committed to in TPP.[18] The change in the estimated STRI from pre-TPP to TPP policy settings is an input to the model.

TPP also obliges countries to accept cross-border services trade obligations on a “negative list” basis, meaning that the signatories promise to provide full access to their services markets unless they specifically list an exception, or NCM reservation. This implies that each TPP partner is making commitments to open trade for the full range of services, except those specifically listed in the NCM annexes. Any new services introduced in the future are also included under TPP’s disciplines. As a result, the negative list approach is likely to be more important to trade in sectors where there are ongoing high levels of innovation. The GTAP services sectors were ranked according to their digital intensity and digital usage in business processes, in order to capture differences in the degree of innovation and likely introduction of new digital services. In the model, it was assumed that the adoption of a negative list approach would reduce barriers to services trade to a greater extent in services sectors which are more digitally intense, as these are likely to be more innovative.[19]

One particular horizontal issue addressed in TPP has gained substantial public attention: the treatment of e-commerce, and specifically cross-border data flows.[20] The ability to manage information efficiently is a critical requirement to keeping down costs of supply in many service sectors. In the model, it was assumed that barriers to services trade were reduced as a result of the helpful disciplines in the TPP E-Commerce chapter. Given that financial institutions and other suppliers of cross-border financial services are excluded from these disciplines, however, it is assumed in the model that the two GTAP financial sectors (ofi and isr) do not benefit from lower barriers to trade from this factor.

Taking into account the liberalization observed in TPP as a result of these three factors, we estimated a combined percentage reduction in observed barriers to U.S. services exports to the other 11 TPP partners. These were expressed as percentage reductions to AVEs reported above.[21] We assumed equal weights for the contribution of each of the three factors to the overall reduction in each AVE, and capped their possible combined contribution at 90 percent. This method expresses changes in barriers to services trade from TPP in terms of relative price changes, which can then be fed into the CGE model, along with assumptions for liberalization in tariffs, quotas, and foreign direct investment (FDI) barriers, to estimate overall income and trade effects.

Table G.7: Percent change in AVEs due to the combined effects of STRI rescoring for TPP, the negative list approach, and TPP e-commerce provisions

	cns	trd	otp+atp	wtp	cmn	ofi	isr	obs	osg
Canada	0.0	-9.8	-0.2	-3.0	-18.8	0.0	0.0	-16.5	-7.5
Mexico	0.0	-9.8	-0.2	-3.0	-18.8	0.0	0.0	-26.0	-7.5
Chile	0.0	-9.8	-0.2	-3.0	-18.8	0.0	0.0	-16.5	-7.5
Peru	0.0	-9.8	-0.2	-3.0	-18.8	0.0	0.0	-16.5	-7.5
Japan	0.0	-24.8	-0.5	-9.0	-45.8	-24.0	-24.0	-40.5	-19.5
Australia	0.0	-9.8	-0.2	-3.0	-18.8	0.0	0.0	-16.5	-7.5
New Zealand	0.0	-24.8	-0.5	-9.0	-45.8	-24.0	-43.5	-40.5	-19.5
Malaysia	0.0	-24.8	-0.5	-9.0	-45.8	-24.0	-24.0	-50.0	-19.5
Singapore	0.0	-9.8	-0.2	-3.0	-18.8	0.0	0.0	-26.0	-7.5
Vietnam	0.0	-24.8	-0.5	-9.0	-45.8	-24.0	-24.0	-40.5	-19.5
Brunei	0.0	-24.8	-0.5	-9.0	-45.8	-24.0	-24.0	-40.5	-19.5

Source: USITC calculations.

Incorporating Investment Provisions

The TPP Agreement would impact the U.S. economy not only by lowering barriers to cross-border trade, but also by reducing barriers to foreign investment. This section describes Commission analysis integrating these foreign investment effects into the dynamic CGE model used to estimate TPP’s effects.

While economists have long recognized the importance of investment to international trade,[22] modeling the investment impact of trade agreements has been difficult. This analysis builds on a model of international investment used in a recent Commission report on trade and investment barriers in India.[23] The current analysis of TPP uses a similar methodology and data, although the changes in investment barriers (or “shocks”) were constructed from a careful consideration of the specific provisions and exclusions in the TPP Agreement.[24] This investment model is not a dynamic model, and key elements of the static model are transferred to the dynamic CGE model used to assess the agreement.

Methodology

The overall goal of this analysis is to calculate the impact of TPP’s investment provisions on economic variables (welfare, employment, etc.) in the United States. However, the Commission does not have a single model capable of doing this, so this analysis goes from the starting point of the TPP text to changes in welfare and productivity in several steps, using an output from one model as an input into the next. The first step is to calculate how much TPP would change investment restrictions, as measured by the OECD’s FDI regulatory restrictiveness index (RRI). Next, the analysis calculates how changes in RRI would affect foreign affiliate sales (FAS) for TPP host countries and foreign affiliate owner countries. Then the analysis calculates how that change in FAS would affect productivity in each sector of each TPP country. Finally, the analysis calculates how those productivity shocks would affect macroeconomic variables in the United States. While the investment model is static, this final step uses the dynamic CGE model.

Figure G.1: Steps in the Commission’s methodology for modeling investment provisions

Source: Compiled by USITC.

Quantifying Changes in Investment Restrictions

The first step is to calculate how much TPP will affect investment restrictions in each TPP member country. This analysis’ measure of investment restrictiveness is the OECD’s FDI Regulatory Restrictiveness Index (RRI).[25] The RRI is a measure of the statutory restrictions on FDI in a particular sector in a particular host country. It is calculated by comparing the host country’s laws concerning FDI in a particular sector to a scoresheet developed by the OECD, with a given restriction on FDI worth a given number of points.[26] The RRI database covers 42 sectors and subsectors in 31 countries in 2014. Table G.8 lists the average RRI for TPP countries. Baseline (pre-TPP) investment restrictions for each sector of TPP countries are measured using the 2014 RRI database.

Table G.8: Investment restrictions (average RRI) in TPP countries

Country	RRI in 2014	RRI after TPP	Change
Australia	0.127	0.112	-0.015
Brunei	0.150	0.130	-0.021
Canada	0.173	0.156	-0.017
Chile	0.057	0.057	0.000
Japan	0.052	0.051	-0.001
Malaysia	0.211	0.139	-0.072
Mexico	0.193	0.170	-0.023
New Zealand	0.240	0.161	-0.079
Peru	0.077	0.070	-0.007
Singapore	0.068	0.053	-0.015
U.S.	0.089	0.074	-0.015
Vietnam	0.150	0.141	-0.010

Source: OECD’s FDI Regulatory Restrictiveness Index (RRI) and USITC calculations.

Notes: RRI values are imputed for Brunei, Singapore, and Vietnam.

While RRI data are available for most TPP countries, they are not available for Singapore, Brunei, or Vietnam. Their RRI values are imputed using the values of similar countries for which RRI data are available. RRI values for sectors in Brunei and Vietnam are imputed using the average RRI value of that sector for all non-OECD member countries in the database. For Singapore, this process is repeated, except that the average of all OECD member countries is used instead.

An additional drawback worth mentioning is that the RRI is a partial measure of the investment climate, not a comprehensive one. It does not include all laws that restrict investment. For example, cultural requirements for TV broadcasting do not factor into RRI.

RRI Changes under TPP

The next step is to calculate how much TPP would change a host country’s RRI in a particular sector. In TPP’s investment chapter, TPP member countries agree not to restrict investment by investors of other TPP countries in certain ways. However, Annexes I, II, and III contain NCMs which specify that TPP’s investment chapter does not apply to certain sectors in certain TPP host countries.

As a result, the Commission splits the calculation of how much TPP will change RRIs into two parts. First, it identified the host country-sectors which have an NCM that partially or fully exempts the country sector from the TPP Investment chapter. Then, for sectors that are not fully exempt, it calculated how much their RRI would fall due to TPP (see Table G.8). Countries, may, of course, reduce their RRI restrictions by more than is required under TPP. But when the Commission’s analysis calculated the level of reform induced by TPP, it assumed that countries would liberalize only the minimum amount required.

The effect of NCMs is to exempt certain sectors from certain provisions of the TPP Investment chapter. In order to calculate the effect of NCMs on RRI, this analysis divided NCMs into two groups: “high” NCMs and “low” NCMs. High NCMs are NCMs that exempt all or almost all of a sector from all or almost all of the TPP Investment chapter. For host country sectors with high NCMs, this analysis assumes that TPP would not change their RRI. Low NCMs are those that exempt only a small part of the sector or exempt a sector only from a small amount of the TPP Investment chapter. For country sectors with low NCMs, the Commission assumes that the RRI would change as much as if there were no NCMs for that country sector at all.[27] Although there are many NCMs that are clearly high, for others the assignment was more subjective.

Next, for sectors not exempted from the TPP Investment chapter by NCMs, the effect of the chapter on RRI was calculated. The TPP Investment chapter’s provisions forbid certain types of investment restrictions, but allow other types. The RRI is scored based on which investment restrictions a country has, out of a specific list of restrictions. TPP forbids all restrictions that compose the RRI except one: restrictions on key foreign personnel. Such restrictions are worth a maximum of 0.1 points of RRI.[28] As a result, for country sectors with a pre-TPP RRI of above 0.1, TPP is assumed to reduce their RRI to 0.1. Country sectors with a pre-TPP RRI of 0.1 or below do not change their RRI.

Table G.9 provides a full list of the projected declines in RRI by host country and sector. For ease of presentation in the table, the change in the index has been multiplied by 100, so that a reported reduction of 6.0 in the table is a change of $-$ 0.06 in the RRI. For instance, $-$ 0.06 is the RRI value for the mining and quarrying sector in Canada, where no high NCMs were identified and the initial RRI would fall from 0.16 to 0.10 due to TPP.

Table G.9: Decrease in RRI from TPP (times 100), by country and sector

Sector	AUS	BRN	CAN	CHL	JPN	MYS	MEX	NZL	PER	SGP	USA	VNM
Agriculture	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
Forestry	0.0	0.0	0.0	0.0	0.0	73.0	0.0	10.0	0.0	0.0	0.0	0.0
Fisheries	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	15.8	0.0	0.0
Mining & quarrying (incl. oil extr.)	0.0	0.0	6.0	0.0	0.0	0.0	0.0	10.0	0.0	0.0	0.0	0.0
Manufacturing	0.0	0.0	1.0	0.0	0.0	0.0	0.3	10.0	0.0	0.0	0.0	0.0
Food and other	0.0	0.0	1.0	0.0	0.0	0.0	0.0	10.0	0.0	0.0	0.0	0.0
Oil refining & chemicals	0.0	0.0	1.0	0.0	0.0	0.0	0.0	10.0	0.0	0.0	0.0	0.0
Metals, machinery, & other minerals	0.0	0.0	1.0	0.0	0.0	0.0	0.0	10.0	0.0	0.0	0.0	0.0
Electric, electronics, & instruments	0.0	0.0	1.0	0.0	0.0	0.0	0.0	10.0	0.0	0.0	0.0	0.0
Transport equipment	0.0	0.0	1.0	0.0	0.0	0.0	1.3	10.0	0.0	0.0	0.0	0.0
Electricity	0.0	0.0	1.0	0.0	0.0	0.0	0.0	10.0	0.0	2.1	14.7	0.0
Electricity generation	0.0	0.0	1.0	0.0	0.0	0.0	0.0	10.0	0.0	2.6	29.3	0.0
Electricity distribution	0.0	0.0	1.0	0.0	0.0	0.0	0.0	10.0	0.0	1.5	0.0	0.0
Construction	0.0	4.2	1.0	0.0	0.0	15.0	0.0	10.0	0.0	0.0	0.0	2.1
Services	0.3	2.2	2.4	0.1	0.2	5.0	4.4	7.3	1.5	0.8	1.8	0.9
Distribution	0.0	4.9	1.0	0.0	0.0	8.5	7.5	10.0	0.0	0.0	0.0	2.5
Wholesale	0.0	2.3	1.0	0.0	0.0	17.0	0.0	10.0	0.0	0.0	0.0	1.2
Retail	0.0	7.5	1.0	0.0	0.0	0.0	15.0	10.0	0.0	0.0	0.0	3.8
Transport	0.0	5.9	0.0	0.0	1.7	6.7	11.1	3.3	0.0	5.4	0.0	0.0
Surface	0.0	0.0	0.0	0.0	0.0	0.0	33.3	10.0	0.0	0.0	0.0	0.0
Maritime	0.0	17.8	0.0	0.0	5.0	20.0	0.0	0.0	0.0	16.2	0.0	0.0
Air	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
Hotels & restaurants	0.0	0.0	1.0	0.0	0.0	0.0	0.0	10.0	0.0	0.0	0.0	0.0
Media	2.5	0.0	20.0	1.0	0.0	22.5	17.5	0.0	15.0	2.5	12.5	0.0
Radio & TV broadcasting	5.0	0.0	20.0	2.0	0.0	25.0	25.0	0.0	30.0	5.0	25.0	0.0
Other media	0.0	0.0	20.0	0.0	0.0	20.0	10.0	0.0	0.0	0.0	0.0	0.0
Communications	0.0	2.3	0.0	0.0	0.0	0.0	0.0	30.0	0.0	0.0	5.0	1.2
Fixed telecoms	0.0	3.4	0.0	0.0	0.0	0.0	0.0	30.0	0.0	0.0	0.0	1.7
Mobile telecoms	0.0	1.2	0.0	0.0	0.0	0.0	0.0	30.0	0.0	0.0	10.0	0.6
Banking	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
Insurance	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
Other finance	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
Business services	0.0	4.0	1.0	0.0	0.0	3.8	0.0	10.0	0.0	0.0	0.0	3.4
Legal	0.0	0.0	1.0	0.0	0.0	15.0	0.0	10.0	0.0	0.0	0.0	5.4
Accounting & auditing	0.0	8.6	1.0	0.0	0.0	0.0	0.0	10.0	0.0	0.0	0.0	4.3
Architectural	0.0	5.2	1.0	0.0	0.0	0.0	0.0	10.0	0.0	0.0	0.0	2.6
Engineering	0.0	2.3	1.0	0.0	0.0	0.0	0.0	10.0	0.0	0.0	0.0	1.2
Real estate investment	30.0	19.3	0.0	0.0	0.0	20.0	6.7	10.0	0.0	6.4	0.0	9.6
Total FDI index	1.5	2.1	1.7	0.0	0.1	7.2	2.3	7.9	0.7	1.5	1.5	1.0

Source: USITC calculations.

Additionally, special treatment is given to certain host countries and sectors. Vietnam’s Annex I NCMs contain an extremely large number of substantial partial exemptions for sectors that are not fully exempt from the investment chapter. In order to deal with this, the Commission’s analysis halves the RRI change for Vietnam. The “TV and radio broadcasting” and “other media” sectors also had many low NCMs. In order to ensure that the RRI changes for these sectors were accurate, their post-TPP RRI includes the effect of both low and high NCM exemptions. Finally, there was no change to RRI in any agricultural sector, because of limited foreign investment in that sector and expected negligible effects based on industry information.

Variation in RRI Changes by Owner Country[29]

The prior section’s calculations show how TPP would change RRI in each host country sector. However, even in a particular host country sector, the change in RRI differs across owner countries. In particular, the United States already has FTAs with a number of TPP countries, and these FTAs already have investment provisions similar to those of TPP. As a result, while TPP would not change the ease of U.S. investment in these TPP countries, it would increase the ease of investment for other TPP members in the aforementioned host countries. However, the OECD provides a single RRI for each host country and sector, for all owner countries.

Table G.10: Change in RRI due to TPP, by owner and host country

Host	Investor
	United States	Other TPP parties with U.S. FTA	Other TPP parties without U.S. FTA	Rest of the world
United States	$-$	No	Yes	No
Other TPP parties with U.S. FTA	No	Yes	Yes	No
Other TPP parties without U.S. FTA	Yes	Yes	Yes	No
Rest of the world	No	No	No	No

Source: Compiled by USITC staff.

Table G.10 shows how the Commission analysis deals with these issues.[30] All countries can be divided into one of four groups: the United States, countries in TPP with which the United States already has an FTA, countries in TPP with which the United States does not already have an FTA, and countries not in TPP. For country pairs marked with a “No,” TPP would lead to no change in RRI, either because it does not apply (for non-TPP countries) or because similar investment provisions are already in place due to preexisting FTAs. For pairs marked with a “Yes,” TPP would lead to changes in RRI; the magnitude of the change in RRI for a particular host country sector would be as calculated in the preceding section.

Continuing the Canadian mining example, the United States already has an FTA with Canada, but Japan does not. As a result, TPP would cause RRI in Canadian mining and quarrying to fall from 0.16 to 0.10 for Japanese investors, but RRI would remain at 0.16 for U.S. investors.

Changes in Foreign Affiliate Sales

Next, the Commission examines how this change in RRI would affect foreign affiliate sales (FAS), which refer to sales by firms located in a domestic market but owned by foreign investors. This analysis uses a database compiled by Commission staff that describes the FAS of each sector, host country, and owner country triplet for 26 host and owner countries and 59 sectors.[31] This analysis also uses econometric analysis from a previous Commission study to describe the relationship between RRI and FAS.[32] The model in that study would predict an increase of 1.8 percent in FAS for each 0.01 decrease in the RRI, holding all else constant.[33] This association is used to estimate FAS changes for each sector, host country, and owner country due to changes in the RRI. For example, as already discussed, the RRI change for mining and quarrying in Canada was $-$ 0.06, which means that FAS in this sector in Canada would increase by 11 percent for all non-US TPP owner countries. However, as explained above, FAS in that sector is not shocked for U.S. or non-TPP owners.

The Commission’s econometric model relies on a number of assumptions. It assumes that the relationship between the restriction index and FAS is similar across sectors. It also assumes that the relationship holds for all host and owner countries. Unfortunately, more detailed data are not available to estimate econometric coefficients that would vary by country. Nonetheless, FAS effects will vary by host country, owner country, and sector, as the RRI varies by host country, owner country, and sector. Additionally, the econometric model for the effect of RRI on FAS does not control for tariff rates.[34] To the extent that FAS are affected by tariffs and tariffs are excluded from the regression and correlated with the RRI, it is possible that the coefficient for the RRI variable may be biased down, and thus the effect of the RRI is overstated.[35] Finally, the econometric model uses the variation in host country RRI that applied to all owner countries. However, a host country reform that only applied to TPP owner countries (and not to non-TPP owners) would increase the consumer price of FAS not owned by TPP countries relative to FAS that are owned by TPP countries. To the extent that the TPP-owned and non-TPP-owned FAS are substitutes, the increase in FAS to owners from countries with a falling RRI may be understated.

Changes in Sectoral Productivity

This section describes how the Commission uses a comparative static CGE model called GTAP-FDI to analyze increases in productivity as a result of changes in FAS due to TPP investment provisions. The GTAP-FDI model is based on the standard GTAP model, extended to include data on FDI and FAS. This model has also been extended to treat the labor force as an endogenous variable (assuming a flexible labor supply). Note that the FDI model uses the same labor supply elasticities as those used in the dynamic GTAP model, which were drawn from the empirical literature.[36]

Under the flexible labor supply assumption, the labor supply elasticity is greater than zero, which implies that the labor supply will expand in response to a rise in real wages, and contract if wages fall. This assumption allows entry into TPP to cause adjustments to aggregate employment in each country.

The simulations use GTAP version 9, with a 2011 baseline. The Commission aggregated 140 regions of the original GTAP model into 26 regions.[37] The 57 GTAP sectors were disaggregated into 59 sectors: retail and wholesale were split into two different sectors, as were telecommunications and other communications.

When a country reduces its restrictions on FDI, costs decrease for the foreign affiliates that it hosts. This leads to increased FAS but also increases the productivity of the host country. This increase in productivity can be calculated from the increase in FAS using the GTAP-FDI model.

This analysis runs 12 simulations using the GTAP-FDI model, one for each TPP member country, in which only that country liberalizes investments that it hosts.[38] In each simulation, that host country’s FAS for all the other 11 owner countries and sectoral productivity parameters are swapped and the host country’s FAS for all owner countries are shocked by the amounts given in the previous step.[39] The GTAP-FDI model then calculates the productivity change in each sector of that country. Table G.11 shows the average productivity gains of TPP member countries from the GTAP-FDI model.

Table G.11: Host country productivity gain from TPP’s reduction in host country RRI, percentage

Country	Productivity gain
Australia	0.075
Brunei	2.202
Canada	0.018
Chile	0.000
Japan	0.001
Malaysia	0.687
Mexico	0.605
New Zealand	0.693
Peru	0.001
Singapore	0.070
United States	0.001
Vietnam	0.021

Source: USITC estimates.

Notes: Economy-wide productivity gains are calculated as share-weighted means of sector/parent gains using sales shares.

As can be seen from table G.11 above, Brunei, New Zealand, and Malaysia would be expected to receive the highest productivity gains as a result of reducing their investment barriers according to TPP investment provisions. This would be due to the fact that these TPP member countries have relatively high initial FDI barriers pre-TPP, and would therefore reduce their FDI barriers more to enjoy higher overall productivity gains. By contrast, countries like Chile, the United States, Japan, and Peru already have fairly low FDI barriers prior to TPP, and therefore have little room to further reduce their FDI barriers based on TPP investment provisions. Hence, the resulting productivity gains for these countries would be relatively low.

Effects on the Economy of Cross-border Trade and FAS

As the final step in its modeling process, the Commission runs a combined simulation incorporating productivity gains as a result of TPP investment provisions, coupled with reductions in tariff and nontariff measures for cross-border trade in goods and services. This last simulation, conducted in the dynamic GTAP model, gives the macroeconomic impacts of the TPP Agreement.

Bibliography

Anderson, James E., and Eric van Wincoop. “Trade Costs.” Journal of Economic Literature 42, no. 3 (2004): 691 $-$ 751.

Borchert, Ingo, Batshur Gootiiz, and Aaditya Mattoo. “Guide to the Services Trade Restrictiveness Database.” World Bank Policy Research Working Paper WPS6108. Washington, DC: World Bank, June 1, 2012. //documents.worldbank.org/curated/en/2012/06/16441094/guide-services-trade-restrictions-database.

$—$ $—$ $—$ . “Policy Barriers to International Trade in Services: Evidence from a New Database.” World Bank Economic Review 28, no. 1 (January 2014): 162 $-$ 88.

Cecchini, Paolo, Michel Catinat, and Alexis Jacquemin. The European Challenge: Benefits of a Single Community. Commission of the European Communities. Brookfield, VT: Gower Publishing, 1988.

Fontagné, Lionel, Amelie Guillin, and Cristina Mitaritonna. “Estimations of Tariff Equivalents for the Services Sectors.” Centre d’Études Prospectives et d’Informations Internationales. Working Paper No. 2011-24. Paris: CEPII, December 2011. www.cepii.fr/PDF_PUB/wp/2011/wp2011-24.pdf.

Fontagné, Lionel, Cristina Mitaritonna, and José Signoret. “Estimated Tariff Equivalents of Services NTMs.” Forthcoming, USITC Office of Economics Working Papers, 2016.

International Labour Organization (ILO). ILOSTAT Database. //www.ilo.org/ilostat/faces/help_home/data_by_subject (accessed March 22, 2016).

International Monetary Fund (IMF). World Economic Outlook databases. //www.imf.org/en/Data#global (accessed February 22, 2016).

International Trade Centre (ITC). TRADE MAP. //www.trademap.org/(X(1)S(wrki5p55djdtwbinhczebb55))/Index.aspx.

International Trade Centre (ITC). Market Access Map (MAcMap). “Tariff Rates for 2014 $-$ 2031 between TPP Member Countries absent the TPP Agreement.” Prepared for the Global Economic Partnership Agreement Research Consortium, 2015.

$—$ $—$ $—$ . “Tariff Rates for 2016 $-$ 2046 between TPP Member Countries under the TPP Agreement.” Prepared for the Global Economic Partnership Agreement Research Consortium, 2016.

Kalinova, Blanka, Angel Palerm, and Stephen Thomsen. “OECD’s FDI Restrictiveness Index: 2010 Update.” OECD Working Papers on International Investment 2010/03. Paris: OECD, 2010. doi:10.1787/112474484663.

Lakatos, Csilla, and Tani Fukui. “The Liberalization of Retail Services in India.” World Development 59 (July 2014): 327 $-$ 40. doi:10.1016/j.worlddev.2014.01.013.

Manyika, James, Sree Ramaswamy, Somesh Khanna, Hugo Sarrazin, Gary Pinkus, Guru Sethupathy, and Andrew Yaffe. Digital America: The Tale of the Haves and the Have-Mores. McKinsey Global Institute report: December 2015. //www.mckinsey.com/industries/high-tech/our-insights/digital-america-a-tale-of-the-haves-and-have-mores.

Organization for Economic Cooperation and Development (OECE). Dataset: Economic Outlook No 95 - May 2014 - Long-term baseline projections. https://data.oecd.org/gdp/real-gdp-forecast.htm (downloaded October 27, 2015).

United Nations. Comtrade database. //comtrade.un.org/.

U.S. International Trade Commission (USITC). Overview of Cuban Imports of Goods and Services and Effects of U.S. Restrictions. USITC Publication no. 4597. Washington, DC: USITC, 2016. //www.usitc.gov/publications/332/pub4597.pdf.

$—$ $—$ $—$ . Hearing transcript in connection with inv. no. TPA-105-001, Trans-Pacific Partnership Agreement: Likely Impact on the U.S. Economy and on Specific Industry Sectors, January 14, 2016.

$—$ $—$ $—$ . Trade, Investment, and Industrial Policies in India: Effects on the U.S. Economy. USITC Publication no. 4501. Washington, DC: USITC, 2014. https://www.usitc.gov/publications/332/pub4501_2.pdf.

$—$ $—$ $—$ . U.S.-Korea Free Trade Agreement: Potential Economy-wide and Selected Sectoral Effects, 2007. USITC Publication no. 3949. https://www.usitc.gov/publications/332/pub3949.pdf.

U.S. Department of Commerce (USDOC). Bureau of Economic Analysis (BEA). “International Transactions,” table 1.1. International Data database. //www.bea.gov/iTable/iTable.cfm?ReqID=62&step=1#reqid=62&step=6&isuri=1&6210=1&6200=1 (accessed February 26, 2016).

$—$ $—$ $—$ . Bureau of Economic Analysis (BEA). “International Services,” table 2.2. International Data database. //www.bea.gov/iTable/iTable.cfm?ReqID=62&step=1#reqid=62&step=7&isuri=1&6210=4&6200=161&6211=165 accessed January 16, 2016.

Washington International Trade Association. Remarks by expert panel at “2015 $-$ 2016 TPP Series Part VI: Services” conference, Washington International Trade Association, Washington, DC, March 9, 2016.

World Trade Organization (WTO), Integrated Data Base (IDB). //tariffdata.wto.org.

World Trade Organization (WTO), International Trade Centre (ITC), and United Nations Conference on Trade and Development (UNCTAD). “Methodology for the Estimation of Non-Ad Valorem Tariffs.” Technical Annex B in World Tariff Profiles 2006. Geneva: WTO, ITC, and UNCTAD, 2006.

[1] See USITC, U.S.-Korea Free Trade Agreement, 2007, Appendix F.

[2] Under perfect competition entering a market is costless which drives the product price down to average cost and reduces profits to zero in the sense that every productive factor receives a wage or a return that is commensurate to its productivity.

[3] While the U.S. trade balance has fluctuated significantly since 1980, its correlation with U.S. GDP is about 0.9, during the same period, which suggests a stable relationship between the trade balance and GDP.

[4] Non-TPP importers of these U.S. products were assumed to consider effective prices which not only reflect changes in market prices but also reflect the quality and technical characteristics of products.

[5] Also referred to as normal trade relations (NTR) rates.

[6] For example, the Japan-Australia Free Trade Agreement will be fully implemented in 2031. 2031 Japan-Australia tariff rates are then used after 2031 in the model.

[7] World Trade Organization (WTO), International Trade Centre (ITC), and United Nations Conference on Trade and Development (UNCTAD), “Methodology for the Estimation of Non-Ad Valorem Tariffs,” 2006, 179.

[8] A fill rate is the rate at which a country's importers use up the quota allocated to them under a TRQ.

[9] International Trade Center, Trade Map; WTO, Integrated Data Base (IDB); UN Statistical Division, Comtrade database.

[10] The conventions described above were used in the conversion of specific and compound tariffs to ad-valorem equivalents and the treatment of lines subject to TRQs

[11] For example, in year 5 of the agreement, the rate on a certain product would be 2 percent under the TPP text. But under a pre-existing free-trade agreement, the rate for that same product would be duty-free. In this case, the rate under the TPP text would be replaced with 0.

[12] In particular, these estimatesoften referred to as the CEPII tariff equivalentsare based on Fontagné, Guillin, and Mitaritonna, “Estimation of Tariff Equivalents,” 2011, as updated in Fontagné, Mitaritonna, and Signoret, “Estimated Tariff Equivalents,” 2016.

[13] Theory suggests an elasticity of 1, although it is often found to vary from that value. Based on past experience, the Commission constrains this parameter to 0.8, but this choice does not affect the results. This treatment essentially divides the left-hand side (log exports) by the GDP of the importer.

[14] Details of the estimations are in Fontagné, Mitaritonna, and Signoret, “Estimated Tariff Equivalents,” 2016.

[15] See Fontagné, Guillin, and Mitaritonna, “Estimation of Tariff Equivalents,” 2011, for details on the derivation.

[16] See Anderson and van Wincoop, “Trade Costs,” 2004.

[17] Several hearing witnesses, industry representatives, and industry groups such as the Coalition of Services Industries have indicated that the three factors listedreduced nonconforming measures (NCMs), negative list treatment, and e-commerce disciplinesexplain the bulk of the likely impact of TPP. NCMs are explained later in this section.

[18] Baseline STRI levels are those published by the World Bank, adjusted by USITC staff for commitments in existing U.S. FTAs. The policy simulation considers changes in baseline STRIs due to commitments in TPP. See Borchert, Gootiiz, and Mattoo, “Guide to the Services Trade Restrictiveness Database,” 2012.

[19] Manyika et al., Digital America: The Tale of the Haves, December 2015, 89.

[20] USITC, hearing transcript, January 14, 2016, 46 (testimony of Peter Allgeier, Coalition of Services Industries).

[21] Fontagné, Mitaritonna, and Signoret, “Estimated Tariff Equivalents,” 2016.

[22] For example, see Cecchini, Catinat, and Jacquemin, The European Challenge, 1988.

[23] USITC, Trade, Investment, and Industrial Policies in India, 2014. Previous Commission studies on prospective FTAs have not included a quantitative assessment of provisions that reduced barriers to FDI.

[24] USITC, Trade, Investment, and Industrial Policies in India, 2014, considers the effects of a hypothetical full removal of FDI restrictions on foreign affiliates in India.

[25] Available at //www.oecd.org/investment/fdiindex.htm. For a description of their methodology, see Kalinova, Palerm, and Thomsen, “OECD ’s FDI Restrictiveness Index,” 2010.

[26] Note that the RRI is a measure of the regulatory restrictions on FDI, not of the regulatory restrictions that apply to all firms. For example, health and safety regulation that apply to all firms do not affect RRI.

[27] This assumption means that the true change in RRI is smaller than what is used in the model.

[28] Kalinova, Palerm, and Thomsen, “OECD's FDI Restrictiveness Index,” 2010, 11.

[29] The Owner Country is the home country of the owners of the investment.

[30] This analysis only includes the effect of U.S. FTAs. Although other TPP countries have bilateral FTAs with each other, their investment provisions may not be the same as those of TPP.

[31] The original database has 140 host and owner countries and 57 sectors. In this simulation, countries are aggregated to 26 regions and sectors disaggregated to 59 sectors.

[32] USITC, Trade, Investment, and Industrial Policies in India, 2014.

[33] See the econometric estimates in appendix G, in USITC, Trade, Investment, and Industrial Policies in India, 2014.

[34] Although tariffs are discussed here, an analogous caveat must also be made for nontariff barriers to importation, which have the same issues.

[35] This would be case, for instance, of “tariff jumping” FDI. At least for the case of tariffs, many of the largest barriers are in the food and agriculture sector, where foreign investment is very low in any case.

[36] Specifically, for the United States and other developed countries in the model, this elasticity is 0.4; for all developing countries, the elasticity used is 0.44.

[37] The 26 regions are the 12 TPP member countries, mainland China, Hong Kong, Indonesia, South Korea, the Philippines, Russia, Taiwan, Thailand, Cambodia, Laos, India, EU, Brazil, and the rest of the world.

[38] This is done because the econometric estimate implicitly assumes a unilateral liberalization. However, the individual unilateral liberalizations are eventually combined into a multilateral liberalization (as in TPP) in the final step with the dynamic GTAP model.

[39] This productivity parameter is country-sector specific.