# Trade and tax reforms in a cash-in-advance economy.

1. IntroductionThe IMF's and World Bank's structural adjustment and stabilization programs often involve a reduction in trade taxes that is accompanied by an increase in other domestic taxes.1 For this reason, much attention has been paid to the welfare, revenue, and market access effects of such trade-tax reforms. The results established in the trade and tax reform literature, noted below, have indubitably generated important policy implications. However, the theoretical work on these issues has so far been conducted exclusively within a barter framework, even though, as is well known, the introduction of money can alter results obtained within a nonmonetary environment. For example, in a Heckscher-Ohlin model with a transactions-based demand for money, Stockman (1985) shows that inflation can cause changes in the pattern of trade even in the absence of changes in real comparative advantage. Also, Kimbrough (1985) shows that tariffs and quotas are not equivalent within a monetary framework because the two policies have different lifetime welfare implications even when they are equivalent in the long run. Similarly, Roldos (1992) demonstrates how the introduction of money modifies the predictions of the specific-factors trade model. Finally, Kemp (1990) and Palivos and Yip (1997a) show, contrary to standard results obtained for barter economies, that an improvement in the terms of trade may not increase the welfare of a small monetary economy. Consequently, free trade may be harmful compared to a no-trade situation.

The purpose of this article is to reexamine some of the issues analyzed in the piecemeal reform literature within a monetary environment. More specifically, we use a model where money must be held in advance for transactions to be consummated to consider isolated tariff reforms as well as two types of joint reform strategies that are very common in the literature: (i) a reduction in import tariffs combined with an increase in consumption taxes so as to keep consumer prices unchanged and (ii) a reduction in export taxes combined with an increase in production taxes so as to leave producer prices unchanged. To preview our results and justify the question posed above, we note that in general the effects of these reforms are different when conducted within a monetary environment and depend on the size of financing frictions in the exportable relative to that in the importable sector. For example, whereas in a barter economy a reduction in tariffs increases welfare unambiguously, we find that in a monetary economy, under certain conditions, such a reduction may actually decrease welfare.

In the existing trade and tax reform literature, the main result has been that reductions in import tariffs (export taxes) combined with increases in consumption (production) taxes improve welfare and government revenue (see, among others, Michael, Hatzipanayotou, and Miller 1993; Hatzipanayotou, Michael, and Miller 1994; Keen and Ligthart 2002; Emran 2005). (2) This occurs because a tariff-tax reform, for example, that leaves consumer prices unchanged improves production efficiency, by reducing the excessive production of the importable goods, and at the same time increases government revenue, by reducing the implicit production subsidies. Likewise, an export and production tax reform that keeps producer prices unaffected improves consumption efficiency, by reducing excessive consumption of the exportable goods, and at the same time increases government revenue, by reducing implicit consumption subsidies.

Market access often plays an important role in trade negotiations. (3) Thus, the recent literature has also analyzed the market access effects of tariff changes. For example, Ju and Krishna (2000) show that tariff reductions that improve welfare may hurt market access. To get a better understanding of this result it is convenient to think in terms of exports. Balanced trade requires that the value of imports equals the value of exports at world prices. A reduction of import tariffs increases the output of the exportable good and thus exports, while a welfare increase (due to lower tariffs) increases the demand for the exportable good and thus reduces exports. Anderson and Neary (2007) and Falvey and Kreickemeier (2009) derive a larger set of welfare or market access-improving reforms in the absence and presence, respectively, of factor market distortions, by using a new approach to the theory of piecemeal trade policy reform. Also, among others, they show that the conflict between reforms that increase welfare and those that increase market access remains valid in their framework. Finally, Kreickemeier and Raimondos-Moller (2008; henceforth KR), in an article that is particularly pertinent to this one, consider the welfare and market access effects of combined tariff-tax reforms and show that such reforms are less efficient in improving welfare and market access than tariff reductions alone. Intuitively, both reforms reduce the rebates given back to the consumers. The tariff-tax reform does it with an increase in a distortionary tax, whereas the reform that uses only tariffs does it in a lump-sum manner. Hence, the latter reform yields a higher welfare. Also, a tariff-tax reform, which is equivalent to a decrease in production subsidies, removes only one reason for importing less, whereas a reduction in tariffs only, which is equivalent to a decrease in consumption taxes and production subsidies, removes two reasons. Consequently, the latter reform results in higher imports.

As mentioned above, the importance of analyzing what may seem as purely international trade issues within a monetary environment has been demonstrated by Kemp (1990) and Palivos and Yip (1997a). The first article uses a money-in-the-utility-function approach, while the second introduces money via a generalized cash-in-advance constraint. (4) Nevertheless, our work differs from these articles in several aspects. First, it does not seek to characterize the optimal tariff or tax, which has been the main concern of the money-trade literature; rather it takes as a starting point the existence of arbitrary levels of distortionary tariffs and/or taxes, as is the case in the trade and tax reform literature. Second, unlike the previous articles, it analyzes situations that involve more than one policy instrument. Third, it is concerned not only with welfare but also with the effects of trade and tax policies on government revenue and market access. Finally, the article complements the tariff and tax reform literature in that it examines similar issues in the presence of liquidity constraints, which are important especially in developing countries.

The remainder of the article is structured as follows. Section 2 presents the theoretical model, which examines trade and tax reforms in the context of a small open monetary economy. Money is introduced via a generalized cash-in-advance (henceforth CIA) constraint, in such a way that cash requirements per unit of value purchased differ across goods. In such a framework, section 3 reexamines the effects of isolated tariff reforms and coordinated tariff-tax reforms on welfare, revenue, and market access. It shows that a uniform radial reduction of tariffs (a reduction of all tariffs by the same proportion) has ambiguous effects not only on market access but also on welfare. Nevertheless, the result that a reduction in import tariffs combined with an equal increase in consumption taxes improves welfare unambiguously is still valid in a monetary economy. The difference between the welfare effects that result from these two types of reforms arises because a CIA constraint introduces a distortion, which may be exacerbated under a tariff-only reform, whereas it remains constant under a coordinated tariff-tax reform. More specifically, a reform that decreases tariffs, and hence consumer prices, affects the total cost of purchasing imported goods relative to the exported. A coordinated tariff-tax reform, on the other hand, does not affect this cost, since it maintains constant consumer prices. It follows then that, contrary to previous results, a tariff-only reform is not necessarily more efficient in improving welfare than a coordinated tariff-tax reform.

Section 4 analyzes the welfare effects of another type of reform strategy, whereby a reduction in the export taxes is combined with an offsetting increase in production taxes so as to keep the producer prices unchanged. (5) In this case, it is shown that, again contrary to previous results, such a reform may lead to a decrease in welfare. The penultimate section of the article analyzes the case where there is endogenous labor-leisure choice and shows that, even if the cash requirement ratios are the same for all importable and exportable goods, some of our previous results are still valid. The reason for this is, of course, the presence of an inherently credit good, leisure, which is nontraded. Hence, the CIA constraint, as we show, generates a wedge between the consumer relative price of leisure and the marginal product of labor. Section 6 concludes with a brief summary.

2. The Model

We develop a model of a small monetary economy based on the barter and money trade models of KR and Palivos and Yip (1997a, b), respectively. Our notation is the same as that in KR except for market access, which is denoted here by A (we reserve M to denote money demand).

Basic Environment

Consider a small open monetary economy that produces and consumes N + 1 tradable goods. Assume that all markets are competitive. Let one good indexed by zero (0) be the numeraire good. Let also [p.sup.w] denote the vector of world prices of the remaining N goods, t > 0 (< 0) be the vector of specific import tariffs (export taxes), and e and [tau] be the vector of production and consumption taxes, respectively. Given that the small open economy cannot influence the world prices, the vector of the domestic producer prices of the non-numeraire goods is given by p = [p.sup.w] + t - [epsilon] = [p.sup.w] - [lambda], where [lambda] = [epsilon] - t is the total tax burden on production. Similarly, the vector of the domestic consumer prices for a small open economy is given by q = [p.sup.w]+ t + [tau] = [p.sup.w]+ [gamma], where [gamma] = t + [tau] is the total tax burden on consumption. There are no taxes applied to the numeraire good; that is, the domestic producer and consumer price of this good, [p.sub.0] and [q.sub.0], are both equal to the world price [p.sup.w.sub.0] = 1. (6)

Consumers maximize their utility function u = u([D.sub.0], ... [D.sub.N]), where [D.sub.i] is the consumption of the ith commodity, i = 0, 1, ..., N. A crucial element of this economy is that a certain fraction of purchases of each good must be financed with cash (see more on the nature of this constraint in the next subsection). More specifically, as in Palivos and Yip (1997a, b), we assume that the following generalized CIA or liquidity constraint holds: (7)

[[phi].sub.x][D.sub.0] + [[phi].sub.m]q'D [less than or equal to] M.

where [[phi].sub.x], [[phi].sub.m], [member of] [0, 1] are positive scalars that denote the cash requirement ratios for purchasing the exportable and importable goods, respectively, and M is money demand. (8)

The consumer's utility maximization problem can be represented in terms of its dual of cost minimization. The expenditure function is defined as (9) [??](1 + [[phi].sub.x], (1 + [[phi].sub.m])q, u) = min{[D.sub.0] + q'D + M:u([D.sub.0], D) = u and [[phi].sub.x][D.sub.0] + [[phi].sub.m]),q' D = M}, where the minimization is performed with respect to the consumption of the numeraire good [D.sub.0], the vector of consumptions of all other goods D. and money demand M, [??](1 + [[phi].sub.x] (1 + [[phi].sub.m])q, u) gives the minimum expenditure that is necessary to achieve a utility level u, given consumer prices q and the CIA constraint. The function [??](.) is concave and linear homogeneous in prices. The latter property allows us to write the expenditure function as E[(1 + [delta])q, u], where E(.) = [??](.)/(1 + [[phi].sub.x]) and [delta] [equivalent to] ([[phi].sub.m] - [[phi].sub.x])/(1 + [[phi].sub.x]). The parameter [delta] measures how much more or less cash per unit of value is required for the purchase of importables relative to the cash required for purchase of the exportable good; hence [delta] captures the proportional increase or decrease in the domestic consumer prices of the importables, depending on whether [[phi].sub.m] is greater or smaller than [[phi].sub.x], owing to the monetary distortion introduced by the CIA constraint. We refer to [q.sub.v] [equivalent to] (1 + [delta])q as the virtual price vector of the importable (non-numeraire) goods.

Note that the asymmetric cash requirements between exportable and importable goods generate a consumption distortion, measured by the size of [[phi].sub.x] relative to that of [[phi].sub.m]. In the special case where [[phi].sub.x] = [[phi].sub.m] = 0 we have a barter economy, where no cash is required for the purchase of a good, whereas when [[phi].sub.x] = [[phi].sub.m] > 0, and hence [delta] = 0. the cash requirement ratios are the same across all goods in the economy. Nevertheless, the most interesting case is the one where [[phi].sub.x] [not equal to] [[phi].sub.m]. For if this is true, then on the one hand, under conditions of free trade where the consumer and producer prices are both equal to world prices (p = q = [p.sub.w]), the marginal rate of transformation between importables and exportables is given by [p.sup.w]. On the other hand. the marginal rate of substitution is given by (1 + [delta])[p.sub.w]. The disparity between these two rates is the main reason behind the nonoptimality of free trade found in Palivos and Yip (1997a).

The first partial derivative of the expenditure function with respect to the ith price, denoted as [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] > 0, yields the compensated demand curve for the ith good. Moreover, the matrix of second partial derivatives [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] is negative definite due to concavity, implying that the compensated demands are negatively sloped. Finally, the first partial of the expenditure function with respect to u, denoted as [E.sub.u], is the reciprocal of the marginal utility of income.

Let Q and R denote, respectively, the (N + 1)- and m-dimensional vectors of outputs and inputs. Assume that there is no international factor movement, implying constant factor endowments. Let also [Q.sub.i] = [F.sub.i]([K.sub.1i], ..., [K,sub.mi]), i = 0, ..., N, be the production function of good i, where [K.sub.ji], j = 1, ..., m denotes the quantity of factor j employed in the production of good i. The revenue function is defined as R(p; [bar.K]) = max{p' Q: [Q.sub.i] [less than or equal to] [F.sub.i]([K.sub.1i], [K.sub.mi]) and [[summation].sup.N + 1.sub.i=1] [K.sub.ji] [less than or equal to] [[bar.K].sub.j] [for all]i , j}, where the maximization is performed with respect to [Q.sub.i] and [K.sub.ji]. Since factor endowments are given, henceforth we suppress K as an argument of the revenue function and write R(p) to denote the maximum value of total output produced by competitive firms that face prices p. given the technology and factor-endowment constraints. The function R(p) is convex and linear homogeneous in prices. The first partial derivative of R(p) with respect to the ith price, denoted as [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], yields the supply of the ith output. Finally, the matrix of second partial derivatives [R.sub.pp] is positive definite due to convexity.

The Cash-in-Advance Constraint

The CIA constraint introduced above, also known as a Clower constraint, is a convenient way to distinguish a monetary from a barter economy. This is so because in a monetary economy commodities are not accepted as means of payments, and money must enter in all or most transactions: "Money buys goods and goods buy money; but goods do not buy goods" (Clower 1967, p. 6). Accordingly, the consumer must hold money in advance (e.g., invest in liquid assets) to finance his or her purchases. Hence, implicitly, the asset market operates before the goods market. Lucas and Stokey (1983, 1987) were the first to introduce the idea that the cash requirement may apply only for a subset of consumption goods. Thus, they distinguish between cash goods, where the cash requirement ratio is one, and credit goods, where the cash requirement ratio is zero. An obvious example of a credit good is leisure. Other authors have assumed that the cash requirement ratio for consumption and/or investment goods is somewhere between zero and one (e.g., Chin, Guo, and Lai 2009). The implication of the CIA constraint is that it increases the cost of a good to the consumer, since a certain amount of money must be held in advance for the purchase to take place. It is as if there is a tax at rate [[phi].sub.i] applied to the particular good i. If the cash requirement ratio is different between two goods, say, 1 and 2, then the existence of the CIA constraint distorts the marginal rate of substitution (MRS) between the two goods, because their ratio of marginal utilities will now be set equal to (1 + [[phi].sub.1])p/(1 + [[phi].sub.2]), where [[phi].sub.i] are the cash requirement ratios and p is the relative price.

In an international context, what is important is the difference between the cash requirement ratios in the exportable and the importable sectors, that say, in terms of our notation, [[phi].sub.x] - [[phi].sub.m]. In general, the exportable and the importable sectors consist of different proportions of durable and nondurable goods. (10) Given that the durable goods are subject to a different degree of credit rationing than the nondurables, we can expect that [[phi].sub.x] [not equal to] [[phi].sub.m]. (11) However, the degree of credit rationing that exists within a country does not depend only on its own credit conditions and financial development, but also on world financial programs. Several countries have established financial institutions, known as Export Credit Agencies (ECAs), that act as intermediaries between governments and exporters/foreign buyers to promote a nation's exports. These institutions provide credits and guarantees either to the exporting firms (seller credits) or directly to foreign buyers (buyer credits). It has been argued that if not for these credits, poor countries would not be able to buy needed imports. (12)

Export credits can be short term (up to two years), medium term (two to five years), and long term (five to ten years). The maturity of the loan depends, among other factors, on the type of product that is being traded. For example, according to the Understandings and Agreements regarding the terms of payments among the Berne Union members, (13)

the maximum length of credit for nondurable consumer goods and consumer services is six months, whereas for consumer durables it is normally six months and maximum two years. It is also explicitly stated that "High unit value does not automatically mean that longer credit should be involved" (Berne Union 2001, p. 5). (14)

In sum, the provision of credit to importers, from foreign governments who seek to promote their exports, makes ceteris paribus the cash requirement ratio for a country's importable goods lower than that for its exportables; that is, [[phi].sub.m] < [[phi].sub.x]. Moreover, the difference between the two ratios increases with the net percentage of durable goods that are imported. (15)

Equilibrium Analysis

The equilibrium in this economy requires that (i) the sum of total spending on goods E(q,,, u) and money holdings M be equal to the sum of income from the production of private goods R(p), the money supply [bar.M], and tax revenue G and (ii) money demand be equal to the money supply. Combining (i) and (ii) yields

E([q.sub.v],u)=R(p)+G. (1)

We assume that the government redistributes its revenue in a lump-sum fashion to domestic households. This government revenue consists of three forms of taxes, namely, trade, consumption, and production taxes; we often refer to the last two forms as "domestic" taxes. Thus,

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)

where [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] denotes the compensated import demand. Market access A is defined as the value of imports at world market prices (see, e.g., Ju and Krishna 2000, Eqn. 3, and KR, Eqn. 6); that is,

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3)

We conclude this section by deriving the effects of changes in domestic consumer and producer prices, q and p, respectively, on government revenue (G), welfare (u), and market access (A). First, take the total differential of Equation 2 with respect to u, q, and p.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (4)

Moreover, differentiating Equation 1 with respect to u, q, p, and G and using Equation 4, we have

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (5)

where [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], assuming that goods are normal in consumption (this is the so-called Hatta normality condition; see Hatta 1977a and 1977b for closed- and open-economy versions, respectively). Finally, differentiation of Equation 3 with respect to u, q, and p results

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (6)

Equations 4, 5, and 6 are the main equations used below to examine the effects of trade-tax reforms on government revenue, welfare, and market access for the small open economy described above.

3. Tariff-Tax Reforms

In this section we examine reforms that involve (i) only tariffs and (ii) tariff and consumption taxes.

Tariff Reforms

Recall that q (= [p.sup.w] + t + [tau]) and p (= [p.sup.w] + t - [epsilon]) denote domestic consumer and producer prices, whereas [gamma] (= t + [tau]) and [lambda] (= [epsilon] - t) are the total tax burdens on consumption and production, respectively. Consider a uniform radial reduction of all import tariffs (t), that is, the so-called proportionality rule, while consumption ([tau]) and production ([epsilon]) taxes are zero. More specifically assume that dt = -[theta]t, where [theta] is a small positive scalar and [tau] = [epsilon] = 0; consequently, dq = dp = dt, [gamma] = t, and [lambda] = -t. Equation 5 then becomes

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (7)

where [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] is negative definite.

Equation 7 shows that a uniform proportional reduction in all tariffs affects welfare through two effects. The first term on the right-hand side (RHS), -[theta]t'St, is positive and denotes the standard welfare effect of a reduction in tariffs. The second term on the RHS, [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], is novel and represents the monetary effect of a radial reduction of imports tariffs on welfare; it is due to the asymmetric cash requirements in the exportable and the importable sectors. In a barter economy, where a financial constraint is absent, that is, [[phi].sub.x] = [[phi].sub.m] = 0 and hence [delta] = 0, a uniform proportional decrease in all tariffs increases welfare unambiguously. This is a standard result in international trade literature within a barter economy context. It is due to the decrease in the production and consumption inefficiencies induced by the decrease in tariffs. The same result holds also in a monetary framework in the special case where [[phi].sub.x] = [[phi].sub.m] > 0 and hence [delta] = 0 again; that is, the cash requirements are the same for the consumption of the exportable and the importable goods.

Nevertheless, in a monetary economy where [[phi].sub.x] [not equal to] [[phi].sub.m] a monetary effect owing to the asymmetric cash requirements is present. The sign of this effect depends on the magnitude of [[phi].sub.x] relative to that of [[phi].sub.m]. In particular, if the exportable good requires more cash balances per unit of value than the importable goods, that is, [[phi].sub.x] > [[phi].sub.m], and hence [delta] < 0, then the monetary distortion affects welfare negatively. Intuitively, if [[phi].sub.x] > [[phi].sub.m], then the additional cost or implicit tax ([[phi].sub.1]) on the exportable imposed by the CIA constraint is higher than that on the importables ([[phi].sub.m]); that is, the virtual relative price of importables, (1 + [delta])q, which is set equal to the MRS, is lower than the market relative price (= q). This will generate excessive consumption of the importable goods, affecting welfare negatively. To offset this negative effect the government should increase the tariff rates; instead, by decreasing the tariff rates, the monetary effect is exacerbated. On the other hand, if [[phi].sub.m] > [[phi].sub.x], then a reduction in tariffs is a movement toward the right direction. Put differently, when [[phi].sub.m] > [[phi].sub.x] and tariff rates decrease, the consumer must now hold exactly [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (the second term on the RHS of Eqn. 7) less money to buy importables relative to that held to buy the exportable good (recall that [delta] is how much more or less money he must hold per unit of value and [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] is the vector of compensated demands for the imported goods). In other words, a reduction in tariffs reduces the price of importables. If the cash requirement ratio in the importable sector is higher than that in the exportable, then the MRS [= (i + [delta])q] is greater than the relative price q. Thus, there is an excessive consumption of the exportable good. The reduction in tariffs will decrease partly this (excessive) consumption of the exportable good and increase the consumption of the importables, increasing thus welfare. We conclude that if [[phi].sub.x] > [[phi].sub.m] ([[phi].sub.m] > [[phi].sub.x]), then the standard efficiency effect and the monetary effect, identified above, work in opposite (the same) directions. Obviously, if [[phi].sub.x] > [[phi].sub.m] and the adverse monetary distortion is sufficiently large to outweigh the efficiency-improving effect, then a uniform proportional cut in tariffs decreases a country's welfare.

Equation 4 can be written as

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (8)

where the second equality follows after substituting away the term -[theta]t'St using Equation 7 (recall also the definition off [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]). The results here are similar to those found in a barter economy; namely, as Equation 8 indicates, a uniform radial reduction of import tariffs has an ambiguous effect on government revenue. The first RHS term of Equation 8, [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], captures the change in tariff revenue due to an income effect. The second RHS term, [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], captures a reduction in tariff revenue, at constant levels of imports, due to lower tariff rates. Finally, the third RHS term, -[theta]t'St, captures an increase in tariff revenue due to a higher level of imports resulting from the radial reduction in tariffs.

Next we examine the effects of a uniform radial reduction of import tariffs on market access. Solving Equation 7 for du and substituting our finding in Equation 6 we obtain

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (9)

where [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] is the marginal propensity to spend on importable goods and is assumed to be strictly between zero and one. Again the results regarding market access are similar to those in a barter economy (see Ju and Krishna 2000). Indeed, if we set [[phi].sub.x] = [[phi].sub.m] = [delta] = 0 in Equation 9, we derive Equation 15 in Ju and Krishna (2000), which in terms of our notation is written as dA = ([p.sup.w] + [beta]t)'Sdt. Accordingly, the effect of an isolated tariff reform has an ambiguous effect on market access, owing to the conflict between a substitution and an income effect, as shown in KR. In a monetary framework, there is an additional (monetary) effect due to the asymmetric cash requirements, which is represented by the second term on the RHS of Equation 9, [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. The sign of this term is also ambiguous since it depends on the sign of [delta]. Its interpretation is the following. Recall that [delta] [equivalent to] ([[phi].sub.m] - [[phi].sub.1])(1 + [[phi].sub.x]) denotes how much more or less cash the consumer spends on importables relative to that spent on the exportable good or, alternatively, the mark-up over the relative prices of importables q faced by a consumer. In other words, the relative cost of importables for the consumer is not just q but rather the virtual price [q.sub.v] [equivalent to] (1 + [delta])q Hence, if [delta] > 0, then the term [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] gives the increase in the income or decrease in the cost to the consumer of buying imports, following a decrease in tariffs (dt < 0). Out of this increase in the income a percentage [beta] will be spent on imports, resulting in an overall increase in imports that is equal to [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. For completely symmetric reasons, the monetary effect is negative on market access when [[phi].sub.x] > [[phi].sub.m] and hence [delta] < 0.

Ju and Krishna (2000) also consider a rule of the form dt = -[delta]([p.sup.w]+ [beta]t) (the "Ju-Krishna rule"). They show that such a rule increases import value for a small open economy. (16) In a monetary environment, however, as can be seen after direct substitution in Equation 9, this reform rule has ambiguous effects on market access, because of the monetary effect identified above. The following proposition summarizes the most important results in this subsection.

PROPOSITION 1. In a small open monetary economy, a uniform proportional reduction of only tariffs has an ambiguous effect on welfare. A sufficient condition for a welfare improvement, following a reduction in tariffs, is that the cash requirement ratio in the exportable sector ([[phi].sub.x]) is lower than that in the importable sector ([[phi].sub.m]). Under the same sufficient condition ([[phi].sub.x] < [[phi].sub.m]), a proportional reduction of tariffs alone according to the "Ju- Krishna rule" raises market access.

Tariff-Tax Reforms

Next we examine the implications of coordinated tariff and consumption tax reforms that leave all consumer prices unchanged, that is, dq = 0, holding production taxes [epsilon] > 0 constant; that is, we make a small generalization relative to the previous literature on piecemeal trade reforms by allowing for production taxes to be present. In the current context of domestic taxes and import tariffs, a radial reduction of all tariffs combined with an increase of all consumption taxes so as to leave the consumer prices unchanged can be represented by dp = dt = [theta][lambda] < 0, where [lambda] = [epsilon] - t is the vector of net production subsidies on the imported goods. Applying dq = 0 and dp = [theta][lambda] to Equation 5 we obtain

[OMEGA]du = [lambda]'[R.sub.pp]dt = [theta][lambda]'[R.sub.pp][lambda] > 0, (10)

where it may be recalled that [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. Furthermore, using Equations 4 and 10 and the fact that in the present context dp= dt= [theta][lambda], we obtain

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (11)

Hatzipanayotou, Michael, and Miller (1994), Keen and Ligthart (2002), and KR have shown, within barter frameworks and with no production taxes, that a coordinated tariff-tax reform increases welfare and government revenue. This occurs because such a tariff-tax reform leaves consumer prices of the importable goods (q) unchanged but reduces their producer prices (p). This improves production efficiency by reducing the excessive production of the importable goods. At the same time, this type of reform increases government revenue, by reducing the implicit production subsidies. Equations 10 and 11 indicate that this result is valid even in the presence of production taxes and asymmetric cash requirements between the exportable and importable goods. Intuitively, since a coordinated tariff-tax reform leaves consumer prices unchanged, the monetary distortion introduced by the asymmetric cash requirements is neutralized. However, when production taxes are present, in addition to import tariffs and consumption taxes, this result holds under the assumption that all imported goods are burdened with a net production subsidy, that is, [lambda] < 0.

Consider next the effect of coordinated tariff-tax reforms on market access. Using Equation 6 and taking into account that in the current context dq = 0 and dp = dt, we obtain

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (12)

Using Equation 10 to substituting for du in Equation 12, we have

dA = ([beta][lambda] - [p.sup.w])] [R.sub.pp]dt. (13)

(Recall that [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] and is between zero and one.) By setting dt = [theta][lambda] in Equation 13 we obtain dA = [theta]([beta][lambda] - [p.sup.w])'[R.sub.pp][lambda], and thus a tariff-tax reform that keeps consumer prices unchanged has an ambiguous effect on market access, regardless of the presence of asymmetric cash requirements. The explanation is once again that this type of reform keeps consumer prices constant, and thus the monetary distortion is neutralized. Hence, we are left with the two conflicting effects that were mentioned in the introduction: namely, a reduction of import tariffs increases the output of the exportable good and thus the value of exports (which is equal to the value imports at world prices), while a welfare increase (due to lower tariffs) increases the demand for the exportable good and thus reduces exports.

To complete this subsection, we consider now the effect on market access of a reduction of tariffs according to the "Ju-Krishna rule" combined with an equal increase of consumption taxes so as to leave consumer prices unchanged. According to the "Ju-Krishna rule" for the coordinated tariff-tax reforms, the reform is of the type dt = [theta]([beta][lambda] - [p.sup.w]), appropriately modified to take into account the presence of production taxes in addition to import tariffs and consumption taxes. By substituting this formula in Equation 13 we obtain dA = [theta]([beta][lambda] [p.sup.w])],[R.sub.pp]([beta][lambda] = [p.sup.w]) > 0, and thus the "Ju-Krishna rule" increases import value in the present framework, under the assumption that all imported goods are burdened with a net production subsidy. We summarize these arguments in the following proposition. (17)

PROPOSITION 2. Consider a small open monetary economy in which there exist tariffs and domestic (consumption and production) taxes. Then a radial reduction of import tariffs combined with an equal increase in consumption taxes that leaves consumer prices unchanged increases welfare and government revenue, if all imported goods are burdened with a net production subsidy. Moreover, under the same assumption regarding the existence of a net production subsidy, a reduction of tariffs according to the "Ju-Krishna rule" accompanied by an equal increase in consumption taxes that keeps consumer prices constant increases market access.

In the next subsection, we compare the welfare and market access effects of a tariff reform alone with those of a coordinated tariff-tax reform. This comparison is useful because it makes obvious the difference between our results and those in KR.

Tariff-Tax Reforms versus Tariff Reforms

We compare the welfare effects of the two reforms in the case where the initial consumption and production taxes are zero. (18) In this case it follows that [gamma] = t and [lambda] = -t. Subtracting Equation 10 from Equation 7 and substituting for [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] we obtain

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (14)

where the second equality follows after setting dt = -[theta]t. Upon setting [[phi].sub.x] = [[phi].sub.m] = [delta] = 0 in Equation 14, the second term on the RHS vanishes, and we obtain the result in KR, namely, that in a barter economy a proportional reduction in tariffs leads to a higher increase in welfare than a proportional coordinated tariff-tax reform that keeps consumer prices unchanged. Intuitively, the coordinated tariff-tax reform replaces tariff revenue with a consumption tax, which is distortionary, whereas a reform that uses only tariffs does it in a lump-sum manner. Hence, the latter reform yields a higher welfare.

The same result regarding the superiority of the tariff-only reform emerges in a monetary economy if [[phi].sub.x] = [[phi].sub.m]) 0, since even in this case [delta] = 0. Hence this extends the results of KR to the case of monetary economy with the same cash requirements for all goods. Nevertheless, if [[phi].sub.x] [not equal to] [[phi].sub.m], there is an additional effect, which is represented by the second term on the RHS of Equation 14, [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. Recall from the previous two subsections that this term emanates from the fact that while a tariff-only reform affects the total cost to consumer of purchasing imported goods relative to the exported, a coordinated tariff-tax reform does not affect this cost, since it maintains constant consumer prices. In particular, if the exportable sector is more liquidity constrained, that is, [[phi].sub.x] > [[phi].sub.m] [??] [delta] < 0, then the term [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] will be negative, and thus the overall effect is ambiguous. Put differently, if the exportable sector requires relatively more cash, then the MRS [= (1 + [delta])q] is less than the relative price q, which implies that there is an excessive consumption of importables. The decrease in tariffs will exacerbate this discrepancy between the MRS and the relative price q, increasing thus the size of the negative effect. On the contrary, in the case of a coordinated tariff-tax reform there is no negative (or positive) monetary effect since consumer prices remain unchanged. Moreover, it follows that if this monetary effect is negative and sufficiently large, then the result in KR that a coordinated proportional tariff-tax reform increases welfare by less than a proportional reduction of only tariffs is reversed. (19)

Next we compare the market access effects of both reforms according to the "Ju-Krishna rule." Subtracting Equation 13 from Equation 9 and setting dt = -[theta]([beta]t + [p.sup.w]) yields

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (15)

KR show that the reform of tariffs alone according to the "Ju-Krishna rule'" increases market access by more than a reduction of tariffs according to the "Ju-Krishna rule" combined with an equal increase in consumption taxes. This can also be seen from Equation 15, where by setting [delta] = 0 we obtain

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

Nevertheless, in a monetary environment, there is an additional effect (captured by the second term on the RHS of Eqn. 15) due to the asymmetric cash requirements. The sign of this second term depends on the sign of [delta], and this renders the overall effect ambiguous. In particular, if the cash requirement ratio in the importable sector is higher than that in the exportable ([delta] > 0), then lowering taxes will decrease the cost of importables relative to that of the exportable good and hence will lead to an increase in imports; in other words, the two effects will reinforce each other. For completely symmetric reasons, if the cash requirement ratio in the importable sector is lower than that in the exportable, then the monetary effect is negative on market access, and hence, overall, the two effects work in opposite directions. We summarize our results in the following proposition.

PROPOSITION 3. In a small open monetary economy, a reform that reduces proportionally all tariffs does not necessarily increase welfare by more than a coordinated tariff-tax reform that keeps consumer prices the same. A sufficient condition for the welfare superiority of tariff-only reforms is that the cash requirement ratio in the exportable sector ([[phi].sub.x]) is lower than that in the importable ([[phi].sub.m]). Under the same sufficient condition ([[phi].sub.x] < [[phi].sub.m]), a reduction of tariffs alone according to the "Ju-Krishna rule" increases market access by more than a proportional reduction of tariffs accompanied by an equal increase in consumption taxes that keeps consumer prices constant.

4. Export and Production Tax Reforms

In this section we examine how a reduction in export taxes with an offsetting increase in production taxes that keeps producer prices unchanged, that is, dp = 0, affects the welfare of a small open monetary economy in an environment where there are export and domestic (production and consumption) taxes. To facilitate the analysis we now assume that the numeraire good is the importable one. The domestic consumer prices of the non-numeraire goods are given by q = [p.sup.w] + t + [tau], where t < 0 is the vector of the export taxes and [tau] denotes as before consumption taxes; hence, once again we make a small generalization of the previous literature by allowing in this case for consumption taxes to be present. On the other hand, the domestic producer prices for the non-numeraire goods are p = [p.sup.w]+ t - [epsilon], where [epsilon] is a vector of production taxes. Note that since we denote export taxes by t < 0, a reduction of their size implies that, algebraically, t rises or that dt > 0. In the present context of domestic taxes and export taxes, a radial reduction of export taxes combined with an equal increase of production taxes, so as to leave producer prices unchanged, is represented by dt = -[theta][gamma], where [gamma] [equivalent to] t + [tau] < 0 is the vector of net consumption subsidies on the exported goods.

The expenditure function is now written as [??][(1 + [[phi].sub.x])q, 1 + [[phi].sub.m], u] = min{q'D + [D.sub.0] + M:u(D, [D.sub.0]) = u and [[phi].sub.x]q'D + [[phi].sub.m][D.sub.0] = M}, where the minimization is performed with respect to the vector of consumption D, the consumption of the numeraire good Do, and money demand M; [??](1 + [[phi].sub.x])q, 1 + [[phi].sub.m], u] gives the minimum expenditure that is necessary to achieve a utility level u, given consumer prices q and the CIA constraint. The function [??](.) is concave and linear homogeneous in prices. The latter property allows us to write the expenditure function as E(1 + [??])q, u), where E(.) = [??](.)/(1 + [[phi].sub.m]) and [??] [equivalent to] ([[phi].sub.x] - [[phi].sub.m])/(1 + [[phi].sub.m]); notice that [??] has the opposite sign from [delta]. In particular, the parameter [??] measures how much more or less cash per unit of value is required for the purchase of exportables relative to the cash required for purchase of the importable good; hence [??] captures the proportional increase or decrease in the domestic consumer prices of the exportables, depending on whether [[phi].sub.x] is greater or smaller than [[phi].sub.m], owing to the monetary distortion introduced by the CIA constraint. The virtual price vector of the exportable (non-numeraire) goods is now [q.sub.v] [equivalent to] (1 + [??])q.

As shown by Keen and Ligthart (2002) and Emran (2005), in a barter economy a reform that involves a reduction in exporting taxes holding producer prices unchanged increases welfare and revenue. Intuitively, this occurs because an export tax is simultaneously a consumption subsidy and a production tax. A reform strategy that lowers export taxes but leaves producer prices unchanged is equivalent to a reduction in the consumption subsidy. Thus, the revenue increases because the cost of the consumption subsidy (indirectly imposed by export taxes) has been reduced. In addition, this reform strategy is welfare enhancing since it improves consumption efficiency by reducing excessive consumption of the exportable goods.

In our framework of a monetary small open economy with a generalized CIA constraint, there exists an additional distortionary effect. In particular, since this reform increases the consumer prices of the exportable goods, the monetary cost of buying these goods rises when [[phi].sub.x] > [[phi].sub.m] or [??] > 0, and hence welfare is affected negatively. More specifically, if there is a higher cash requirement ratio in the exportable sector, then a decrease in export taxes raises the consumer price of exportables, which means that the monetary cost of buying exportables goes up (recall that the consumer must hold a fraction [[phi].sub.x] of the value of purchases calculated at domestic prices). Thus, the overall welfare effect is ambiguous. If, however, [[phi].sub.x] < [[phi].sub.m] and hence [??] < 0, then the monetary effect strengthens the aforementioned efficiency effect; thus, welfare rises unambiguously following a reduction in export taxes. More formally, to examine how a radial reduction of all export taxes combined with an equal increase in production taxes that keeps producer prices unaffected changes welfare and government revenue in the presence of a CIA constraint, set dp = 0 and dq = dt = -[theta][gamma] in the equation that is analogous to Equation 5 to get

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (16)

Equation 16 indicates that a reduction in export taxes with an offsetting increase in production taxes, so that producer prices remain unchanged, has an ambiguous effect on welfare in the presence of asymmetric cash requirements. The first term on the RHS of Equation 16 denotes the standard direct effect that affects welfare positively, if all exported goods are burdened with a net consumption subsidy, that is, [gamma] < 0. The second term on the RHS denotes the indirect effect on welfare due to difference in cash requirements between exportable and importable goods. The sign of this term is ambiguous and depends on the size of [[phi].sub.x] relative to that of [[phi].sub.m]. in particular, if the cash requirement ratio for purchasing the exportable goods is higher than the cash requirement for purchasing the importable good, that is, if [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], and all exported goods are burdened with a net consumption subsidy, then this second term is negative. Intuitively, the rise in the price of exportables, following a reduction in export taxes, will increase the amount of money that the consumer must hold in order to buy the exportable goods. Notice that this additional amount of money is given by the term [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. which appears on the RHS of Equation 16. In sum, if the cash requirement ratio in the exportable sector is higher than that in the importable, then the overall welfare effect depends on the relative strength of the two opposing effects. Obviously, if [[phi].sub.x] > [[phi].sub.m] and the adverse monetary distortion is sufficiently large to outweigh the efficiency-improving effect, then a uniform proportional cut in export taxes combined with offsetting raises in production taxes so that all producer prices remain unchanged will decrease the country's welfare.

Next, applying dp = 0 and dq = dt = -[theta][gamma] to the equation that is analogous to Equation 4 and using Equation 16, we obtain

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (17)

Equation 17 indicates that this type of reform raises government revenue even in the presence of asymmetric cash requirements. The reason is, of course, that it reduces the consumption subsidies indirectly imposed by export taxes. However, in the current context of consumption taxes in addition to export and production taxes, a sufficient condition for this type of reform to increase government revenue is that all exported goods are burdened with a net consumption subsidy.

PROPOSITION 4. Consider a small open monetary economy in which there exist export and domestic (consumption and production) taxes. Then, a radial reduction in export taxes combined with an offsetting increase in production taxes, so that all producer prices remain unchanged, has an ambiguous effect on welfare. Sufficient conditions for an increase in welfare are (i) that the cash requirement ratio in the exportable sector is lower than that in the importable and (ii) all exported goods are burdened with a net consumption subsidy. Under sufficient condition (ii), this reform increases government revenue.

The results derived so far in the literature depend critically on the assumption that there are no financial constraints in the economy. Our results show that, when financial constraints are taken into account, then the reduction of all export taxes combined with an equal increase in production taxes so as to leave producer prices unchanged may reduce welfare. Thus, the existence of "win-win" reform strategies in a monetary economy depends crucially on the nature of the CIA constraint for purchasing goods.

5. Endogenous Labor-Leisure Choice

In this section we show that our results can be extended to the case of endogenous labor-leisure choice. Not only is this case interesting in its own right, but it is also a case where the distinction between cash and credit goods arises naturally, since leisure is inherently a credit good. For the sake of brevity, we analyze only the case of a reduction in the tariff when there are only two tradeable goods, an exportable, which is indexed with "0" and is used as the numeraire good, and an importable, which is indexed with "1."

Let l and L denote leisure and labor supply, respectively. The time endowment is normalized to one, so that l + L = 1. The domestic relative (consumer and producer) price of the imported good is q = p = [p.sup.w] + t, where [p.sup.w] is as before the world price of the imported good and t denotes a specific import tariff. Consumers maximize their utility function, which depends on the consumption of the two tradeable goods and leisure, u = u([D.sub.0], [D.sub.1], l) = u([D.sub.0], [D.sub.1], 1 - L), where [D.sub.0], [D.sub.1], and l are all assumed to be normal goods. A notable difference between this and the previous sections is that here we assume that money is equally efficacious in each market, that is, [[phi].sub.x] = [[phi].sub.m] = [phi] [member of] [0, 1]. The CIA constraint is now written as

[phi][D.sub.0] + [phi]q[D.sub.1] [less than or equal to] M. (18)

The consumer's expenditure function is now defined as: [??][1 + [phi], (1 + [phi])q, L, u] = min{[D.sub.0] + q[D.sub.1] + M:u([D.sub.0], [D.sub.1], 1 - L) = u and [phi][D.sub.0] + [phi]q[D.sub.1] [less than or equal to] M}, where the minimization is again performed with respect to [D.sub.0], [D.sub.1] and M; [??][1 + [phi], (1 + [phi])q, L, u] gives the minimum expenditure needed to achieve a utility level u, given consumer prices q, employment L, and the CIA constraint. The function [??](.) is linear homogeneous in prices. Thus, we can write the expenditure function as E(q, L, u), where E(.) = [??](.)/(1 + [phi]). The expenditure function is increasing in the relative price q and in L, strictly concave in q (i.e., [E.sub.qq] < 0) and strictly convex in L (i.e., [E.sub.LL] > 0). Its derivative with respect to q ([E.sub.q]) gives the compensated demand function for the imported good. On the other hand, its derivative with respect to L ([E.sub.L]) gives the reservation wage in terms of the numeraire good [??]/(1 + [phi]). The latter is the amount of addition to income that will elicit further supply of labor (see Dixit and Norman 1980). The cross-partial derivative [E.sub.Lq], captures the way the imported good and leisure are related in consumption. If [E.sub.Lq] > 0, then they are substitutes, since an increase in the price of importables will increase the reservation wage and hence will increase the demand for leisure (decrease the supply of labor). Similarly, if [E.sub.Lq] < 0, then the importable good and leisure are complement goods. Finally, since all tradeable goods and leisure are normal, [E.sub.qu] > 0 and [E.sub.Lu] > 0.

The revenue function is defined again as R(p, L; [??] = max{[Q.sub.0] + p[Q.sub.1]:[[summation].sub.i][L.sub.i] [less than or equal to] L, [Q.sub.i] [less than or equal to] [F.sub.i] ([L.sub.i], [K.sub.li], ... [K.sub.m - li]), and [[summation].sub.i] [K.sub.ji] [less than or equal to] [[bar.K].sub.j] [for all]i, j}, where [Q.sub.i] denotes the output of good i = 0, 1, [K.sub.ji] ([L.sub.i])j = 1, ... m - 1 denotes the quantity of factor j (labor) employed in the production of good i, and the maximization is performed with respect to [Q.sub.i], [L.sub.i], and [K.sub.ji]. We assume that the production functions of both goods F(.) exhibit constant returns to scale. Henceforth, we omit the fixed factors from the revenue function and write R(p, L) to denote the maximum value of total output produced by competitive firms that face prices p, given the technology, the labor supply, and factor-endowment constraints. The first partial derivative of R(.) with respect to p, [R.sub.p], yields the supply of the imported good and with respect to L, [R.sub.L], the marginal revenue product of labor. Moreover, the function R(p, L) is strictly convex in p and strictly concave in L, that is, [R.sub.pp] > 0 and [R.sub.LL] < 0. Finally, the cross-partial derivative [R.sub.pL] constitutes a general equilibrium measure of factor intensity of the imported good (see Dixit and Norman 1980). If [R.sub.pL] is positive (negative), then the domestic output of the imported good increases (decreases), following an increase in labor supply. Thus, if [R.sub.pL] > 0 (< 0), then we call the imported good labor intensive (nonlabor intensive).

As before, we assume that the government redistributes all its revenue in a lump-sum fashion to domestic households. The equilibrium in this economy requires that (i) total spending on goods E(q, L, u) and money holdings M be equal to the income from production of private goods R(p, L), plus the money supply [bar.M], plus tariff revenue [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], (ii) the money demand be equal to money supply, and (iii) the reservation wage be equal to the marginal revenue product or the market wage rate. Combining the first two conditions yields

E(q, L, u) = R(p, L) + t([E.sub.q] - [R.sub.p]). (19)

Condition (iii), on the other hand, means that

(1 + [phi]) [E.sub.L] (q, L, u) = [R.sub.L] (p, L). (20)

Next, we examine how a reduction of the import tariff will affect the welfare of the economy. Totally differentiating Equations 19 and 20 with respect to u, L, q, p, and t and using the fact that dq = dp = dt we obtain

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (21)

where it may be recalled that [OMEGA] [equivalent to] [E.sub.u] - [tE.sub.uq] > 0. First, notice that when labor supply is exogenous, the welfare effect of a reduction in the tariff can be found immediately from Equation 19 to be

du / dt = t([E.sub.qq] - [R.sub.pp]) / [OMEGA] < 0,

since [E.sub.qq] < 0, [R.sub.pp] > 0; that is, a reduction in the tariff increases welfare. This is a standard effect whereby a reduction in the tariff increases consumption and production efficiency and hence welfare. The same result holds when labor supply is endogenous. To see this, denote the determinant of the LHS coefficient matrix in Equation 21 by [DELTA]. It follows that [DELTA] > 0 for stability reasons (see Hatzipanayotou and Michael 1995). Then, in a barter economy, where [phi] = 0,

dU / dt = T[([E.sub.qq] - [R.sub.pp])([E.sub.LL] - [R.sub.LL] - [([E.sub.qL] - [R.sub.pL]).sup.2]] / [OMEGA] < 0,

since [E.sub.qq] < 0, [R.sub.pp] > 0, [E.sub.LL] > 0, and [R.sub.LL] < 0. Thus, a reduction m the import tariff again increases welfare. Nevertheless, in the case of a monetary economy, this result is not necessarily valid. Indeed, if [phi] [not equal to] 0, then it follows from Equation 21 that

du / dt = t([E.sub.qq] - [R.sub.pp])[(1 + [phi])[E.sub.LL] - [R.sub.LL]] - [(1 + [phi])[E.sub.qL] - [R.sub.pL]] [[phi][E.sub.L] + t([E.sub.qL] - [R.sub.pL])] / [DELTA], (22)

whose sign is in general ambiguous. Intuitively, since the CIA constraint is applied to the consumption of the two tradeable goods but not to the consumption of leisure, the MRS between leisure and the consumption of a good is not equal to the relative price. For example, the MRS between leisure and the importable good is equal to w/[(1 + [phi])q], where w is the wage rate, whereas in a barter economy it is equal to w/q. A reduction in the tariff rate lowers q and hence increases the relative price of leisure in terms of the importable. This implies a decrease in the consumption of leisure and an increase in labor supply. The decrease in the consumption of leisure affects also the consumption of the importable good in a way that depends on whether the two goods are substitutes or complements; as mentioned above this is determined by the sign of [E.sub.qL]. The increase in labor supply, on the other hand, will affect the supply of importables in a way that depends on whether the importable good is labor intensive or not; this is determined in turn by the sign of [R.sub.pL].

6. Conclusions

A voluminous theoretical literature examines the welfare and revenue effects of coordinated trade-tax reforms. Recently considerable attention has been paid also to the market access effects of reform strategies. Kreickemeier and Raimondos-Moller (2008) have shown, within a barter economy, that coordinated tariff-tax reforms are less efficient to improve market access and welfare than reforms that involve only tariffs.

In this article we have extended the analysis of Kreickemeier and Raimondos-Moller (2008) to the case of a monetary economy. We have shown that the existence of a financial constraint weakens and may reverse the results in KR. Also, if the exportable goods are more liquidity constrained, then, contrary to previous results, tariff-tax reforms that leave consumer prices unchanged may increase welfare and market access by more than the reforms of only tariffs. Furthermore, in the presence again of a financial constraint, a reform strategy that lowers export taxes but leaves producer prices unchanged may be less desirable or even undesirable if the exportable goods are more liquidity constrained than the importable. Finally, we have extended our results to the case of endogenous labor-leisure choice, where the distinction between cash and credit goods arises naturally.

References

Anderson, James E., and J. Peter Neary. 2007. Welfare versus market access: The implications of tariff structure for tariff reform. Journal of International Economics 71:187-205.

Berne Union. 2001. The Berne Union general understanding. Accessed December 8, 2009. Available at http://www.nexi. go.jp/insurance/ins_berun/pdf/berun1.pdf.

Chao, Chi-Chur, and Chong K. Yip. 2000. Urban unemployment and optimal trade policy in a Cash-in-Advance Economy. Journal of Economics 71:59-77.

Chao, Chi-Chur, and Chong K. Yip. 2001. Non-traded goods and optimal trade policy in a Cash-in-Advance Economy. Journal of International Trade and Economic Development 10:23-37.

Chao, Chi-Chur, and Eden S. H. Yu. 1999. Shadow prices and trade restrictions in a monetary economy. Journal of Macroeconomies 21:755-64.

Chin, Chi-Ting, Jang-Ting Guo, and Ching-Chong Lai. 2009. Macroeconomic (in)stability under real interest rate targeting. Journal of Economic Dynamics and Control 33:1631-38.

Clower, Robert W. 1967. A reconsideration of the microfoundations of monetary theory. Western Economic Journal 6:1-9.

Dixit, K. Avinash, and Victor D. Norman. 1980. Theory of international trade. Cambridge, UK: Cambridge University Press.

Eaton, Jonathan. 1988. Credit policy and international competition, in Strategic trade policy and the new international economics, edited by Paul Krugman. Cambridge, MA: MIT Press, pp. 115-45.

Emran, M. Shahe. 2005. Revenue-increasing and welfare-enhancing reform of taxes on exports. Journal of Development Economics 77:277-92.

Emran, M. Shahe, and Joseph E. Stiglitz. 2003. Price-neutral tax reform with an informal economy. Accessed June 5, 2009. Available at http://econwpa.wustl.edu.

Emran, M. Shahe, and Joseph E. Stiglitz. 2005. On selective indirect tax reform in developing countries. Journal of Public Economics 89:599-623.

Engel, Charles, and Jian Wang. 2007. International trade in durable goods: Understanding volatility, cyclicality, and elasticities. Working Paper No. 3. Dallas: Federal Reserve Bank of Dallas, Globalization and Monetary Policy Institute.

Falvey, Rod, and Udo Kreickemeier. 2009. Tariff reforms with rigid wages. Economic Theory 41:23-39.

Gianturco, E. Delio. 2001. Export credit agencies: The unsung giants of international trade and finance. Westport, CT: Quorum Books.

Hatta, Tatsuo. 1977a. A theory of piecemeal policy recommendations. Review of Economic Studies 44:1-2l.

Hatta, Tatsuo. 1977b. A recommendation for a better tariff structure. Econometrica 45:1859-69.

Hatzipanayotou, Panos, and Michael S. Michael. 1995. Tariffs, quotas, and voluntary export restraints with endogenous labor supply. Journal of Economics 62:185-201.

Hatzipanayotou, Panos, Michael S. Michael, and Stephen M. Miller. 1994. Win-win indirect tax reform. Economics Letters 44:147-51.

IMF. 2005. Dealing with the revenue consequences of trade reform. Background Paper for Review of Fund Work on Trade. Washington DC: IMF Fiscal Affairs Department.

Ju, Jiandong, and Kala Krishna. 2000. Welfare and market access effects of piecemeal tariff reform. Journal of International Economics 5l-52:305-16.

Keen, Michael, and Jenny E. Ligthart. 2002. Coordinating tariff reduction and domestic tax reform. Journal of International Economics 56:489-507.

Kemp, C. Murray. 1990. The gains from trade for a monetary economy. Kobe Economic and Business Review 35:27-30.

Kimbrough, P. Kent. 1985. Tariffs, quotas and welfare in a monetary economy. Journal of International Economics 19:257-77.

Kreickemeier, Udo, and Pascalis Raimondos-Moller. 2008. Tariff tax reforms and market access. Journal of Development Economics 87:85-91.

Lucas, Robert E., Jr., and Nancy Stokey, 1983, Optimal fiscal and monetary policy in an economy without capital. Journal of Monetary Economics 12:55-93.

Lucas, Robert E., Jr., and Nancy Stokey, 1987. Money and interest in a cash-in-advance economy. Econometrica 55:491-514.

Michael, S. Michael, Panos Hatzipanayotou, and Stephen M. Miller. 1993. Integrated reforms of tariffs and consumption taxes. Journal of Public Economics 52:417-28.

Palivos, Theodore, and Chong K. Yip. 1997a. The gains from trade for a monetary economy once again. Canadian Journal of Economics 30:208-23.

Palivos, Theodore, and Chong K. Yip. 1997b. The effects of import quotas on national welfare: Does money matter? Southern Economic Journal 63:751-60.

Palivos, Theodore, and Chong K. Yip. 2006. Optimal intervention policies for a small monetary economy. Journal of Economics 88:69-85.

Rajaram, Anand. 1994. Tariff and tax reform: Do World Bank recommendations integrate revenue and protection objectives? Economic Studies Quarterly 45:321-38.

Roldos, E. Jorge. 1992. A dynamic specific-factors model with money. Canadian Journal of Economies 25:729-42.

Stockman, C. Alan. 1985. Effects of inflation on the pattern of international trade. Canadian Journal of Economies 18:587-601.

(1) See, for example, Rajaram (1994) kind IMF (2005).

(2) A notable exception is Emran and Stiglitz (2005), who show that in the presence of an informal sector coordinated trade-tax reforms may reduce welfare.

(3) Market access is defined as the value of imports at world prices.

(4) Other articles that analyze trade issues in monetary environments include Palivos and Yip (1997b), Chao and Yu (1999), Chao and Yip (2000, 2001), and Palivos and Yip (2006), to name but a few.

(5) See Keen and Ligthart (2002), Emran and Stiglitz (2003), and Emran (2005) for this type of reform.

(6) Henceforth, we omit the price of the numeraire good as an argument of all functions defined below. Note also that the numeraire good can be an importable good (see, for example, Emran 2005). In fact, for reasons that will become transparent later, in section 4 we assume that the non-numeraire goods are exported, while the numeraire good is imported.

(7) A prime denotes transposition of a vector.

(8) It should be emphasized that the results remain the same if one uses other ways to introduce money, such as the money-in-the-utility function (MIUF) or the transactions cost approach. We chose the CIA approach in its present form because it highlights the economic forces that drive the results. For a reexamination of the optimality of free trade using the MIUF approach see Kemp (1990).

(9) As is common in the literature, we assume that the CIA constraint is strictly binding and hence in equilibrium holds with equality.

(10) For example, in the United States in 2007 exports of durable goods constituted 68.2% of total exports of goods. On the other hand, imports of durable goods constituted 59.1% of total imports of goods (Bureau of Economic Analysis, NIPA accounts). For other economies, the difference between these two shares can be even bigger: for example, in Australia, Iceland, and New Zealand exports (imports) of durable goods as a share of total exports (imports) in the year 2000 were 0.25 (0.64), 0.28 (0.51), and 0.22 (0.58), respectively (see Engel and Wang 2007, table 4).

(11) In 2007 the shares of debt of U.S. families were, for home purchase, 69.5%; for home improvements, 2.3%: for other residential property, 10.8%: for investments other than real estate, 1.6%; for vehicles, 5.7%; for other goods and services (durable and nondurable), 6.2%; for education, 3.6%; and for other unclassified purposes, 0.5% (Federal Reserve Board, 2007 Survey of Consumer Finances).

(12) For a discussion of the arguments for and against export credits see Eaton (1988).

(13) The Berne Union, or officially the International Union of Credit and Investment Insurance, is the leading association for export credit and investment insurance worldwide. It was established in 1934 and currently has 54 members in 43 countries.

(14) Calculations by First Washington Associates in the year 2000 indicated that exports covered by all of the world's ECAs (Berne Union and non-Berne Union) amounted to about 12% of total global exports. Some developed country ECAs support an even higher percentage of national exports, and most developing country ECAs support a lower percentage (see Gianturco 2001).

(15) Palivos and Yip (1997b) regard the assumption [[phi].sub.x] = [[phi].sub.m] = [phi] restrictive for one additional reason, namely, that if that is the case. then the velocity of money, defined as [[D.sub.0] + q'D]/[bar.M] = [[D.sub.0] + q'D]/[([[phi].sub.x][D.sub.0] + [[phi].sub.m]q'D] = 1/[phi], is constant and in particular independent of prices, which contradicts the existing empirical evidence.

(16) Note than even though the "Ju-Krishna rule'" increases market access, it cannot ensure welfare improvement. This result can be derived in terms of our notation by substituting dt = -[theta]([p.sup.w] + [beta]t) in Equation 7 and setting [delta] = 0; moreover, it remains valid in the current monetary framework as well.

(17) KR extend the "Ju-Krishna rule" in order to examine how a reduction in tariffs according to this rule affects market access, when it is accompanied by an equal increase in consumption taxes (the "modified Ju-Krishna rule"). They show that the "modified Ju-Krishna rule" cannot ensure an increase in welfare. The same result obtains in the current framework as well.

(18) We do so because the comparison is meaningful when two reforms start from the same equilibrium, hence the same taxes. Moreover, it is possible to draw a conclusion, without further assumptions, only in the case where initial taxes are zero (see also KR, note 9).

(19) One may think that it is obvious that having both a tax and tariff reform versus just a tariff reform can do better, because there are more degrees of freedom. Nevertheless, this is not so for, as mentioned above, in the barter economy case a tariff-only reform leads to higher welfare than a combined tariff-tax reform. Even in the monetary economy, the result is in general ambiguous. Actually, there are not more degrees of freedom in a tariff-tax reform than there are in a tariff only reform, since in the former case the two instruments, the tariff t and the consumption tax [tau], do not move freely. Rather, they have to satisfy the constraint dq = dt + d[tau] = 0. We thank an anonymous referee for insightful questions on this issue.

Theodore Palivos, Department of Economics, University of Macedonia, Salonica 540 06, Greece: E-mail tpalivos@uom.gr.

Nikos Tsakiris, Department of Economics, University of Ioannina, Ioannina 451 10, Greece: E-mail ntsak@cc.uoi.gr; corresponding author.

The authors would like to thank Panos Hatzipanayotou. Leo Michelis, kind two anonymous referees for valuable comments and suggestions on an earlier draft of the article.

Received July 2009: accepted March 2010.

Printer friendly Cite/link Email Feedback | |

Comment: | Trade and tax reforms in a cash-in-advance economy. |
---|---|

Author: | Palivos, Theodore; Tsakiris, Nikos |

Publication: | Southern Economic Journal |

Geographic Code: | 1USA |

Date: | Apr 1, 2011 |

Words: | 11385 |

Previous Article: | Weathering the storm: measuring household willingness-to-pay for risk-reduction in post-Katrina New Orleans. |

Next Article: | National and international tests of the new drug cost offset theory. |

Topics: |