Role of capital controls and exchange rate regime in China: a calibrated
DSGE approach
Xinyi Zhao1
Master Thesis Monetary Policy and Banking Track Amsterdam School of Economics, University of Amsterdam
Supervisor: Dr. Marcelo Zouain Pedroni
July 2016
1
Statement of Originality
This document is written by Xinyi Zhao who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Role of capital controls and exchange rate regime in China: a calibrated DSGE approach
Xinyi Zhao
Abstract: This paper uses a small open economy New Keynesian model with characteristics of China’s policy framework included in the benchmark model. It examines the impluse response of a positive foreign interest rate shock on the domestic economy in terms of stability in real GDP and inflation. It shows that when capital controls are higher, both the level effects and substitution effects contribute to smaller movements in real GDP and inflation. However, the immediate benefits from higher capital control are very limited compared to the damaging effects in the long run. In addition, the paper finds that managed exchange rule that is external looking does a better job in stabilizing the economy than both fixed exchange rate regime and a managed exchange rate that is inward looking. The findings suggest further easing in cross border capital flows conditions and conducting a managed exchange rate to shield domestic economy from positive foreign interest rate shocks.
3 1. Introduction
With pegged nominal exchange rate, strict controls on the cross border capital mobility and sterilization, monetary policy framework of China is much different than most other advanced economies, and is different from assumptions such as complete asset market in standard open economy monetary policy models. The parallel evolution of large current account surpluses and
substantial accumulation of foreign reserves illustrates an interesting feature of the Chinese economy. On the one hand, there are strict restrictions on private capital inflows, which characterize a closed economy. On the other hand, there are substantial net capital outflows, through the accumulation of international reserves. This hybrid system differs from the usual open economy or closed economy paradigms and has received little attention in the literature (Philippe, Kenza, & Yannick, 2013). To analyze the macroeconomic behavior of the Chinese economy, it seems fundamental to have a good understanding of this specific structure.
Moreover, the domestic and global landscapes have been changing continually in the past decade. Since the mid of 2014, China’s foreign reserves have been decreasing at a rapid pace that is unprecedented. At the beginning of 2016, the decline reaches 100 billion per month. Consequently, China has lost nearly 20% of its value in foreign reserves compared to the peak of 4 trillion. In line with this, in the year 2015 alone, the exchange rate has declined by 5.7% against the US dollar (USD). Below the superficial phenomenon lays the fundamental reasons that are two-folded. First, from the domestic perspective, growth has been slowing down. To support a flagging economy, the central bank cuts policy interest rates and reserve requirements in a series of small steps. Moreover, in an effort to promote CNY internationalization, which includes joining well-known international financial benchmarks, such as the Financial Times Stock Exchange (FTSE) and Morgan Stanley Capital International (MSCI) equity indexes and the IMF’s Special Drawing Rights (SDR), policymakers have been easing obstacles to cross border capital flows. This mix of easier monetary policy and more porous capital controls has generated the pressure on currency depreciation. Second, from the external perspective, expectations of a hike in US interest rate have further intensified the situation. After pushing through a landmark rise of 0.25% in short-term interest rates in December 2015, US Federal Reserve policymakers further open up the option of lifting short-term interest rates at their next meeting in June 2016. As China has long been anchoring the nominal exchange rate against the USD, it is threatened by potentially large capital outflows caused by this event.
4
The above landscape may have important implications for China’s policymakers on adjusting the pace of moving towards a more market-oriented economy. Under the background of a potentially foreign positive shock (in this case particular to the US interest rate shock), it is interesting to investigate whether or not China should slow down its policy reform towards less capital controls and more flexible exchange rate regime for the merits of economic stability.
To answer this question, the paper compares the effects of different degrees of capital controls on domestic economy in response to a positive foreign interest rate shock under the pegged exchange rate regime to investigate whether or not China should strengthen its current degree of capital controls. Moreover, to reflect the policy movements to a more market-oriented exchange rate regime, the paper also compares pegged exchange rate regime with managed exchange rate given certain capital controls to investigate whether or not China should slow down its pace on moving towards a more flexible exchange rate regime. The approach this paper applied to investigate the questions is the small open economy dynamic stochastic generalized equilibrium (DSGE) paradigm.
Some literatures find that under certain circumstances, a semi-open economy with imperfect capital mobility can be welfare enhancing since that the semi-open economy also enables inter-temporal trade, which is the main reason that an open economy typically produces a higher welfare than a closed economy. Moreover, in a semi-open economy the central bank can choose a real interest rate different from the world interest rate. In this context, Philippe, Kenza, & Yannick (2013) argue that, with the assumption that consumers are credit-constrained, the domestic interest rate on savings should be temporality above the international rate and that there should be more foreign asset accumulation than in an open economy for a rapidly growing economy like China. Using a different DSGE model that features incomplete market with time varying tax on capital inflows, imperfect asset substitutability that inhibits international risk sharing and nominal rigidities, Liu & Spiegel (2015) also shows that with a combination of capital account restrictions and sterilization of capital flows, a central bank can mitigate the effects of excess foreign capital flows caused by external shocks.
Yet, few papers pay special attention to the role of a combination of a pegged exchange rate and capital controls in terms of shielding domestic economy from the external shocks. With a pegged exchange rate alone, the central bank loses the discretionary monetary policy. Consequently, domestic interest rate has to move one on one to any movement in the foreign
5
interest rate. Foreign shocks have full spillovers in this case. And in a monetary model with nominal rigidities, real variables can be severely affected as well. However, when a pegged exchange rate is used in combination with capital controls, the central bank regains the discretionary monetary policy power, which can be used to stabilize the economy. With capital controls, domestic interest rate does not need to move one on one to movements in the foreign interest rate. Moreover, capital outflows are restricted, which has the potential in restricting the effects of external shocks on domestic economy.
An exception is Chang, Liu, & Spiegel (2015), who design an open economy DSGE model that is specially tailored to China’s characteristics of nominal exchange rates peg, sterilization and capital controls. They introduce a trade-off between fiscal sterilization costs and price stability for the Ramsey central bank that maximizes household utility. Consequently they find that, in response to the negative foreign interest rate and export demand shocks, relaxing both the capital controls and pegged exchange rate regime could improve China’s welfare during the financial crisis.
This paper differs from Chang, Liu, & Spiegel (2015) from the following perspectives. First, with the intention to address the current landscape that China is facing risks of a hike in the US interest rate, the model is examined against positive foreign interest rate shock instead of a negative one. According to Schmitt-Grohe & Uribe (2016), who develop a open economy model with wage rigidity and pegged exchange rate, optimal capital controls should be countercyclical, which means that the government tends to restrict capital inflows when the country is hit by positive shocks. Second, instead of investigating the optimal policy with a Ramsey central bank, the model assumes policy rules and pays attention to stability issues. Third, the paper examines the role of managed exchange rate regime as an alternative exchange rate regime instead of flexible exchange rate to reflect a more relevant policy alternative for China.
The paper finds that when capital controls are higher, both the level effects and
inter-temporal substitution effects contribute to smaller movements in real GDP and inflation. However, the immediate benefits from higher capital control are very limited while the long run effects of having higher capital controls seem to be more damaging. With a subjective discount factor higher than 1, it is more likely that the damaging long term effects dominate the short term benefits of having higher capital controls. On the exchange rate regime, the paper finds that managed exchange rate is better than the fixed exchange rate by largely reducing the variations
6 in real GDP. However, this is true only when the managed exchange rate responds to the external economy rather than the domestic economy. In other words, the central bank should leave the task of stabilizing domestic inflation to interest rate while using the exchange rate as an instrument dealing with external imbalance caused by the shock.
The remainder of the paper is structured as follows. Section 2 gives a short overview of literature on the open economy DSGE models. Section 3 introduces characteristics of China’s monetary policy framework, which are important to be included in the model explained in Section 4. Section 5 presents the calibration. Section 6 shows the role of capital controls and exchange rate regimes, with the dynamic of the model presented in Section 7. Section 8 wraps up the paper and concludes.
2. Literature review on open economy DSGE models
Much recent work in macroeconomics has involved the development and evaluation of monetary models that bring imperfect competition and nominal rigidities into the DSGE structure. The resulting models, where monetary settings generally have non-trivial effects on real variables, are often referred to as New Keynesian. This model consists of an IS curve relating aggregate output to expected future output and the real interest rate, and an augmented Philips curve, in which current inflation is a function of expected future inflation and real marginal cost. Real marginal cost is a function of output relative to the flexible price equilibrium level of output. The model is closed by assuming that the monetary authority determined the nominal interest rate (Walsh, 2010).
Examples of classic literature on the standard open economy New Keynesian model includes Galí & Monacelli (2005), who consider infinite number of countries with Calvo-pricing type nominal rigidities, Walsh (2010), who considers a two-country model with Calvo-pricing type nominal rigidities, and Rogoff & Obstfeld (1996), who build a two-country model with preset price. Besides, more assumptions on nominal rigidities and market frictions are assumed and are incorporated into the model to better capture the reality. One example is Adolfson, Laséen, Lindé, & Villani (2005), who develop a DSGE model for an open economy and estimate it on euro area data using Bayesian estimation techniques. The model incorporates a number of nominal and real frictions including incomplete exchange rate pass-through in both import and export market, sticky wages, variable capital utilization, capital adjustment costs and habit
7
persistence.
This paper adds to the current open economy New Keynesian DSGE literature in the sense that it investigates particularly the effects of capital controls and different exchange rate regimes on the external interest rate shocks’ transmission channel. In a standard open economy New Keynesian model, portfolio adjustment between domestic and foreign bonds is inconsequential because there are perfect substitution and complete asset market with unconditional interest rate parity (UIP) always holding. However, when capital controls exist, asset market is incomplete with imperfect substitution between domestic and foreign bonds. Representing capital controls by portfolio adjustment costs, this means the relative demand for foreign bond will affect pricing of domestic currency, which will further transmit to domestic economy because of the existence of nominal rigidity. Moreover, portfolio adjustment costs also affect consumer’s inter-temporal decisions, as it will offset some of the benefits of holding bonds. In addition, the existence of portfolio adjustment gives rise to the room for exchange rate policy. Compared to fixed exchange rate regime, managed exchange rate regime has the potential to complement capital controls by directly affect the pricing of domestic currency. All these mechanisms require thorough investigation. This paper is aimed to fill this gap.
This paper’s finding on the choice of exchange rate regime also accords to the literature on the optimal monetary policy in open economies DSGE model that considers incomplete asset market. For example, Corsetti, Dedola, & Leduc (2010) study the optimal monetary stabilization policy in interdependent open economies by proposing a unified analytical framework that systematizes the existing literature. They find that when there is incomplete asset market, failure to internalize international monetary spillovers results in attempts to manipulate international relative prices to raise national welfare, causing inefficient real exchange rate fluctuations. Thus they claim that the targeting rules should be not only in output gaps and inflation, but also in misalignments in the terms of trade and real exchange rates, and cross-country demand imbalances, which is in line with the finding in this paper that exchange rate regime shall be used to counter the financial imbalances caused by the external interest rate shocks.
The model in the paper is based on a simplified two-country small open economy model with nominal rigidities, adjusted by a few characteristics featuring China’s policy framework according to Chang, Liu, & Spiegel (2015). Before introducing the model, a basic introduction on the China’s policy framework is introduced in the next section.
8 3. Development of China’s monetary policy framework
This section is aimed at providing background and current stylized facts on the framework of China’s monetary policies that are important to be considered in the model setting. The development of exchange rate system, capital account policy and monetary policy of China are described. Specifically, four characteristics stand out: (1) China’s exchange rate has been strictly managed against USD since 2005; (2) with nominal exchange rate as the primary nominal anchor, sterilization is an important monetary policy that affect the quantitative supply of money; (3) China has moved a couple of steps on improving its interest rate transmission channel; (4) China has pervasive capital controls on portfolio investment while direct investments are significantly less regulated. Recently, some partial liberalization on capital account has been introduced.
3.1. China’s exchange rate regime
In 1994, China transformed its nominal exchange rate regime from a dual-track exchange rate system into a semi-pegged regime. China’s nominal exchange rate was virtually fixed against the USD between 1996 and 2005. Since 2005, the CNY was revalued upwards (appreciated) against the USD by about 2% and the currency was thereafter managed against an undisclosed basket of currencies. In practice, the CNY remained tightly managing against the USD, with the PBOC subsequently indicating that on any given trading day the bilateral exchange rate would be allowed to float within a band of 0.3% relative to the closing price of the previous day. The CNY was re-pegged to the dollar during the global financial crisis and then once again allowed to appreciate against the USD from June 2010. In April 2012, the floating band was widened to 1% per day. Since around the beginning of 2015, China’s exchange rate ends its appreciation trend to depreciation trend until now. The dynamic movement of nominal exchange rate and real exchange rate as a result of policy change can be seen in Figure A1.
3.2. China’s monetary policy
According to the Law of the People’s Republic of China on the PBOC enacted in 1995, the aim of monetary policies is to maintain the stability of the value of currency and thereby promote economic growth (Article 3). Thus, the main nominal anchor for China’s monetary policy has been the nominal exchange rate. China’s monetary instruments include reserve requirement ratio,
9
central bank base interest rate, rediscounting, central bank lending, open market operation and other policy instruments specified by the State Council2, among which sterilization through open market operations has been an important component. As both the current account and the capital account have large surpluses, the PBOC has purchased substantial amounts of foreign currencies in order to meet its nominal exchange rate anchor. Consequently, foreign reserves accumulated rapidly, reaching its peak of 4 trillion in 2014 (see Figure A2). To finance the purchases of these foreign assets, PBOC can either sell domestic bonds or print more money. The former approach is called sterilization. Since 2003, the PBOC has been issuing substantial amounts of central bank bills (CBB) to sterilize the effects of foreign market operations on domestic money supply.
In China, the central bank imposes an upper bound on deposit rates and a lower bound on lending rates. The tight regulation preserves the big four large banks and prevents smaller banks from entering the market. The consequence is that the interest rate channel of monetary transmission is relatively weak because big banks have a broad base of retail deposits that tend to be less sensitive to changes in baseline rates (Prasad & Zhang, 2014). This explains why the PBOC has tended to rely on quantity measures such as sterilization mentioned above. Yet since the late 1990s, some of the restrictions have been relaxed to improve the interest rate transmission channel. In 2012, the upper/lower bound was relaxed to 10 percent above/below the benchmark rate. In July 2013, the PBOC scrapped the floor on lending rates, allowing banks to compete in offering cheap loans to attract the best projects (Song, Storesletten, & Zilibotti, 2014). In 2015, the ceiling on bank deposit interest rates was removed. Finally, Figure A3 shows the benchmark rates of 3-month deposit and lending rate in China, compared with the US Treasury bill rate. It should be noted that China’s interest rates have been higher than the US Treasury bill rate since the financial crisis in 2008, the magnitude of the difference, however, has been reduced.
3.3. Capital control policy
The “impossible trinity”, which compels policymakers to choose only two of three from capital mobility, discretionary monetary policy, and a fixed exchange rate, implies that China has to have its capital account strictly regulated if it does not want to lose control over domestic monetary policy. This is true in reality, where controls are exercised mainly by restricting
2
10
international portfolio investments in China (Song, Storesletten, & Zilibotti, 2014). Portfolio inflows and outflows of China are rather small. According to Song, Storesletten, & Zilibotti (2014), the inward and outward portfolio positions and China’s direct investments abroad (outward direct investments) are an order of magnitude smaller than their counterpart figures in countries with open capital accounts. An exception is that foreign direct investment into China (inward direct investments) is not far from the average level in the countries with open capital accounts.
Nevertheless, a partial liberalization of capital account has taken place over the past decade. For instance, since 2002, the Chinese Securities Regulatory Committee has allowed “qualified foreign institutional investors” (QFIIs) to buy Chinese stocks and bonds. In addition, although nonbank Chinese residents and institutions are still banned from purchasing foreign securities directly, the government has softened the restriction since 2006 by allowing “qualified domestic institutional investors” (QDIIs) to invest in foreign capital markets.
4. Benchmark model
The model used in this paper is a two-country small open economy DSGE model according to Chang, Liu, & Spiegel (2015), which captures specific Chinese charateristics explained in the previous section. Specifically, it considers (1) a pegged exchange rate regime; (2) central bank’s flow-of-fund constraint to reflect the role of government intervention in the foreign exchange rate market; (3) an incomplete asset market where foreign investors are not allowed to hold domestic bond and domestic investors have portfolio adjustment costs when adjusting their portfolio; and (4) price rigidities and imperfect competition so that monetary disturbances have real effects.
4.1. The representative household
The economy is populated by a continuum of infinitely lived households. The representative household maximizes the lifetime expected utility function specified by:
E! β![lnC!+ A!ln !!! ! − A! !!!!!! !!!! ] ! !!! ,
where the period’s utility depends positively on the aggregate real consumption of a bundle of domestic and imported foreign good C!, positively on the holding of real money balance in terms
11 of domestic production price !!
!!, and negatively on labor supply L!. The parameter β is the
subjective discount factor; A! and A! are the utility weights for real money balances and leisure
respectively; σ! is the inverse Frisch elasticity of labor supply.
The maximization is subjected to the budget constraint given below, according to which the representative household chooses the allocation of resources between real consumption, real money balance, and holdings of domestic bonds and foreign bonds denominated in foreign currency: C!+!!! ! + !!!!!!!"∗ !! 1 + !! ! !! !!!!!!!"∗ − ψ ! ≤!!!! !! + !!!! !! + !!!!!!!!!!!!!!!∗ !!,!!!∗ !! + !! !! , (1) where W! denotes nominal wage; R! and R∗! denote the nominal gross interest rates for domestic
bonds B! and foreign privately holding bonds B!"∗ respectively; S
! is the nominal exchange rate;
Π! denotes the nominal dividends received by the household from its ownership of domestic
firms.
According to Schmitt-Grohé & Uribe (2003), when there is incomplete asset markets in the small open economy, the equilibrium dynamics possess a random walk component. To resolve the problem, they propose three remedies, namely a model with an endogenous discount factor, a model with a debt-elastic interest-rate premium and a model with convex portfolio adjustment costs. Following Chang, Liu, & Spiegel (2015), I include the convex portfolio adjustment costs function in the budget constraint, of which the size of the portfolio adjustment cost is represented by the parameter Ω!. The use of the portfolio adjustment costs can be interpreted as in part bureaucratic barriers erected by the China’s government to restrict private agent’s access to foreign assets. The convex function means that any deviation from the steady state portfolio share of domestic bonds in the total value of private bond holdings ψ incurs losses to household. However, the quadratic form of the portfolio adjustment cost function means that its first order approximation is linear and monotonous. As will be shown later, only marginal increases in either domestic bond or private bond holding will incur marginal costs to household.
With Λ! denoting the Lagrangian multiplier for the budget constraint and ψ! representing for !!
!!!!!!!"∗ , the first order conditions for household are given by
w.r.t C!: !!
12 w.r.t m!: !!!!! !!! = 1 − E! !!!!! !! ! !!!!! , (3) w.r.t L!: !!! ! = !!!!!! !! , (4) w.r.t B!: E!!!!!!! ! !! !!!!= 1 + !! ! ψ!− ψ !+ (1 − ψ!) Ω!(ψ!− ψ), (5) w.r.t B!"∗ : E !!!!!!! ! !!∗!!!! !!!!!! = 1 + !! ! ψ!− ψ !− ψ !Ω!(ψ!− ψ), (6)
Combing (2) and (5) gives the Euler equation: E! !!! !!!! !! !!!!= 1 + !! ! ψ!− ψ !+ (1 − ψ!) Ω!(ψ!− ψ). (7)
This condition states that if the household chooses to hold an additional unit of domestic bond, current consumption will drop by one unit plus the marginal portfolio adjustment costs. The next period consumption will increase by the real payoffs from the one unit holding of domestic bond. At the optimum, the marginal benefit of a unit holding of domestic bond must equal its marginal costs. Note that because the right hand side of equation (7) can be re-written as 1 +!!
! ψ!−
ψ (2 − ψ!− ψ), and because ψ! < 1 always holds, only increase in marginal holdings of domestic bond will be costly. By contrast, decline in marginal holdings of domestic bond actually decreases the marginal cost of reducing consumption today.
From (2) and (3), household’s intra-temporal optimization between money demand and consumption is !!!!!! !! = 1 − E! !!! !!!! ! !!!! . (8)
From (2) and (4), the optimal labor supply and consumption is
!!
!! = A!L!
!!C
!. (9)
Combining (2), (5) and (6), household’s intra-temporal optimization between domestic bond and foreign bond is E! !!! !!!! ! !!!! R!− !!∗!!!! !! = Ω!(ψ!− ψ), (10)
which is a generalized form of UIP condition. In the absent of adjustment costs i.e., Ω!= 0, (10) is reduced to R!− R∗!E!!!!!!
! = 0, which is identical to the standard UIP condition that links
relative interest rates to the expected change in the nominal exchange rate. With portfolio adjustment costs, domestic and foreign bonds are imperfect substitutes. The equation represents a downward-sloping curve for relative demand for domestic bonds: when the relative price of
13 domestic bonds falls, the household’s optimal share of domestic bond holdings increases a finite amount.
4.2. Domestic Firms
The domestic firms’ sector consists of three types of agencies, namely the final good firm, the importing firm and the intermediate good firm. The first two types of agencies operate on competitive markets while the third type of agency operates on a monopolistic market.
4.2.1. The representative final goods firm
Final good firms use the whole set of differentiated intermediate goods as inputs and produce homogeneous final consumption good, which is either consumed by domestic households, used by the importing firm as inputs, or exported. With aggregation technology, the production function takes the form of Dixit-Stiglitz aggregator:
Y! = Y!,! !!!! !! di ! ! !! !!!! , θ! > 1 (11)
where Y!,! is the output of intermediate firm i; and θ! is the elasticity of substitution between
intermediate goods.
Given output price P! and intermediate good price P!,!, the final good firm solves the profit
maximization problem: max !!,! P! Y!,! !!!! !! di ! ! !! !!!! − !P!,!Y!,! di !
The solution gives the demand function for intermediate goods Y!,!:
Y!,!= !!!
!,!
!!
Y!. (12)
Introducing (12) into (11), one gets the expression for domestic goods price index:
P! = !P!,!!!!!di
!
!!!!!!
14 4.2.2. Firms in the importing sector
Firms in the importing sector combine domestically produced final goods Γ!,! and imported
goods Γ!,! to produce Γ!, which will be used as inputs for the intermediate good sector. The
production function is specified as below: Γ! = Γ!,!!Γ
!,!!!!, 0 < α < 1. (14)
Given the price for importer’s production P!! and input price P
! for domestic goods and S!P!∗ for
the foreign goods, firms in the importing sector solves the cost minimization problem: min !!,!!,!P! !Γ ! = P!Γ!,! + S!P!∗Γ!,! s. t Γ! = Γ!,!! Γ !,!!!!.
Combing the first order conditions with respect to Γ!,! and Γ!,!, and expressing Q! =!!!!∗
!! as real
exchange rate, one gets
Q! =!!!! (!!,!
!!,!). (15)
Equation implies that the relative demand for the home produced final goods is positively related to the real exchange rate. In other words, real depreciation in domestic currency is associated with higher relative demand for home produced inputs. Introducing (15) into the cost function, and denoting the relative price Q!,! =!!!
!! gives the relationship between real exchange rate and
the relative price Q!,!:
Q!,! = Q!!!! !
!!
!!!!!!. (16)
4.2.3. The monopolistically competitive intermediate goods firm
There is a continuum of intermediate goods producers who are monopoly suppliers of its own differentiated good with labor and output from the importing sector as inputs. Both input markets are competitive. The production function of the intermediate good firm i is given by:
Y!,!= Γ!,!! Z!L!,! !!!, (17)
where Z! is technology that grows at a constant rate λ! ( !!!
!!!= λ!). The real marginal cost mc!
15 minimizing production cost W!L!,!+ P!!Γ
!,! with prices of final product and inputs as given.
Combing the first order conditions with respect to L!,! and Γ!,!, one gets
!! !!!!!!,!= !!! ! !!,! !!,! . (18)
Real marginal costs can be further expressed by mc! = !!
!!!!
!!!
Q!!,!ϕ, where ϕ = ϕ!! 1 − ϕ !!!. (19)
Calvo (1983) and Rotemberg (1982) are the two most popular pricing schemes in the New-Keynesian business cycle literature (Ascari, Castelnuovo, & Rossi, 2011). Following Liu & Spiegel (2015), I incorporate nominal rigidities assuming that the intermediate goods firm faces a quadratic price adjustment cost function given by3
!! ! !!,! !!!,!!!− 1 ! Y!, (20)
where Ω! measures the size of price adjustment costs and π is the steady state inflation rate. Intermediate goods firms solve the profit maximization problem given the price level, the demand schedule for its product, and input prices. Specifically, it solves
max !!,! E! β!C ! C!!![ P!,! P!!!Y!,!!!− mc!!!Y!,!!! − Ω! 2 P!,!!! πP!,!!!!!− 1 ! Y!!!] ! !!! s. t. Y!,!!! = P!!! P!,! !! Y!!!
By symmetry, all firms who re-optimize the price will set the same price. Thus, the subscript i can be ignored and the solution implies the following first order condition price setting rule:
mc! = !!!! !! + !! !! [ !! ! − 1 !! ! − βE! !!!! ! − 1 !!!! ! !! !! !!!! !!!! ], (21)
It can be seen from (21) that when there is no price adjustment costs i.e. Ω! = 0, real marginal cost equals the inverse markup i.e. mc! =!!!!!
! .
3
Here I followthe original work from Rotemberg (1982) who uses units of aggregate ouput to normalize adjustment costs instead of following the version by Chang, Liu, & Spiegel (2015) who for the convenience of calculation normalize adjustment costs in units of aggregate consumption.
16 4.3. The foreign economy
Following Chang, Liu, & Spiegel (2015), I assume that the foreign economy is not allowed to hold domestic bonds because of capital controls in the home country. This incomplete assest market assumption breaks the international sharing mechanism that links foreign consumption with domestic consumption. Thus, the foreign economy is assumed to be exogenous. It follows that foreign consumption and interest rates are given by
lnC!∗ = lnC∗, (22)
lnR∗! = 1 − ρ
! lnR∗+ ρ!lnR∗!!!+ e!!,!, (23)
where C∗ and R∗ are the steady-state levels of the foreign interest rate and consumption respectively; ρ! represents the persistence of the foreign interest rate shock; and ϵ!,! is the shock that follow a normal process with a mean of zero and a standard deviation of σ!.
Finally, the foreign demand X! for the domestic consumption good is defined as negatively
related to the real exchange rate:
X!= Q!!!!C
! ∗Z
!, (24)
where the export demand is augmented by the level of domestic technology to obtain balanced growth.
4.4. Equilibrium
4.4.1. Current account and balance of payments
The balance of payments implies that the current account balance equals the net foreign capital outflows:
CA!= S!B!∗− S!B!!!∗ , (25)
CA! denotes the capital account, which equals net export and net interest income received from holdings of foreign assets:
CA!= P!X!− S!P!∗Γ!,!+ S!B!!!∗ (R∗!!!− 1), (26)
4.4.2. Central bank’s flow-of-fund constraint
The total foreign bond holding is determined by equation (25), the central bank has to make up the residuals between the total foreign bond holding and private foreign bond holding demand by households by purchasing or selling government holding of foreign bond. The central bank purchases (sells) foreign currency revenues from (to) exporters (importers) at the prevailing
17 exchange rate, which is financed by either adapting money supply or domestic bonds. The process is expressed by the flow-of-fund constraint:
B!!− R
!!!B!!!! + M!!− M!!!! = S!(B!,!∗ − R∗!!!B!,!!!∗ ). (27)
where B!! and M !
! denote the supply of domestic bond and money respectively, and B
!,!∗ is the
government’s holding of the foreign bond. The aggregate capital outflow B!∗ is:
B!,!∗ + B
!,!∗ = B!∗. (28)
4.4.3. Market clearing
Final goods market clearing implies that aggregate output is used by domestic consumption, exports and inputs for the importing sector, and is used to compensate adjustment costs:
C!+ X!+ Γ!,!+!! ! !! ! − 1 ! Y!+!!!!!!!,!∗ !! !! ! ψ!− ψ ! = Y !. (29)
The expression can also be derived from combining the household budget constraint (2), central bank flow-of-fund constraint (27), and the external balance equations (25), (26) and (28).
The market clearing conditions for labor, intermediate goods, money and domestic bond markets are: L! = !!L!,! di, (30) Γ! = Γ!,! di ! ! (31) M! = M!!, (32)
Finally, define real GDP as the sum of consumption and net exports:
GDP!= C!+ X!− Q!Γ!,! (34)
4.5. The nonlinear system of equations
This section summarizes the nonlinear system of equations explained before with all variables expressed in real terms. Moreover, because of the constant technology growth, a number of variables are non-stationary, as they contain a real stochastic trend. Thus, I detrend all the real variables by c! =!!! !, y! = !! !!, γ!,! = !!,! !! , γ!,! = !!,! !! , γ! = !! !!, gdp! = !"#!
!! and all the nominal
18 variables by: w!= !!! !!!, m! = !! !!!! , b! = !! !!!!, b! ∗ = !!∗ !!!!∗, b!,! ∗ = !!,!∗ !!!!∗, b!,! ∗ = !!,!∗ !!!!∗, ca! = !!! !!!! .
Further, I represent the nominal depreciation rate by λ!,! = !!
!!!! .
From the household utility maximization, there is Euler equation (35), money demand (36), labor supply (37), and relative demand for domestic bond holdings (generalized UIP condition) (38). E! !!! !!!! ! !! !! !!!!= 1 + !! ! ψ!− ψ !+ (1 − ψ!) Ω!(ψ!− ψ), (35) !!!! !! = 1 − E! !!! !!!! ! !! ! !!!! , (36) w!= A!L!!!c !, (37) E! !!! !!!! ! !! ! !!!! R!− R! ∗λ !,!!! = Ω!(ψ!− ψ). (38)
From the importing firm, cost minimization gives the optimal relative demand for domestic goods inputs (39) and the relationship between relative price and real exchange rate (40). Equation (41) is the production function for the importing firm. Equation (42) gives the definition of real exchange rate.
Q! =!!!! (!!,! !!,!), (39) Q!,! = Q!!!! !!! !!! !!!, (40) Γ! = Γ!,!! Γ !,!!!! (41) Q! = !!!!!!,!!! ∗ !! . (42)
From the intermediate goods firms’ cost minimization and profit maximization, there are relative demand for labor (43), real marginal cost determination (44) and price setting equation (45). From the final goods firms’ profit maximization accompanied by labor market and intermediate goods market clearing conditions, the aggregate production function is expressed by (46).
!! !!,! = !!! ! !! !! , (43) mc! = ϕ w! !!!Q !,! ! , (44) mc! = !!!! !! + !! !![ !! ! − 1 !! ! − βE! !!!! ! − 1 !!!! ! !! !! !!!! !!!!], (45) y!= γ!!L!!!! . (46)
19 In the external sector, export demand is given by (47). Total foreign bond holding is pinned down by the balance of payments (48) and (49). Central bank buys foreign bond and finances it by either printing money or issuing domestic bonds (50).
x! = Q!!!C ! ∗. (47) ! !!ca! = b! ∗−!!!!∗ !!∗!!, (48) ! !!ca! = ! !!x!− γ!,!+ !!!!∗ !!∗! !(R!!! ∗ − 1), (49) b!!−!!!!!!!!! !!!! + m! !−!!!!! !!!! = Q!(b!,! ∗ −!!!!∗ !!,!!!∗ !!∗!! ). (50)
Finally, good market (51), money market (52) and the domestic bond (53) and foreign bond (54) markets all clear. The definition of GDP is given by (57).
c!+ x!+ γ!,!+!!! !!!− 1 ! y!+ (b!+ Q!b!,!∗ )!!! ψ!− ψ ! = y!, (51) m! = m!!, (52) b!,!∗ + b !,! ∗ = b ! ∗, (54) gdp!= c!+ x!− Q!γ!,!. (55)
Thus, with all foreign variables {C!∗, R ! ∗, π
!
∗} exogenously given, an equilibrium in this
economy consists of an allocation of {c!, γ!,!, γ!,!, γ!, x!, m!, m!!, L
!, w!, mc!, b!, b!!, b!,!∗ , b!,!∗ , b!∗, y!,
ca!, gdp!} and (relative) prices {π!, Q!,!, Q!} that satisfy the 21 equations (35) - (55) listed above plus one foreign exchange rate policy and one monetary rule specified in different scenarios that determine { λ!,!, R!}.
5. Calibration
A summarization of the calibrated parameters can be seen in Table 1. The presentation of how the steady state is calculated and a summarization of targeting values can be found in Appendix 2. The data used in the following calibration for targeting is from the time period 2005Q1 to 2015Q4. The reason of choosing only data after 2005 is to reflect the fact that China ends its prolonged period of fixed exchange rate regime since 2005. The real GDP per capita, import share of GDP are from World Bank’s World Development Indicators. Data on inflation, exchange rate and money aggregate (M2) are from International Monetary Fund’s (IMF) International Financial Statistics. Data on domestic government bond (in USD) is from Bank for
20 International Settlements (BIS) Debt Securities Statistics. And data on balance of payments is from IMF’s Balance of Payments Statistics. Short-term interest rate in China and US are from Organisation for Economic Co-operation and Development (OECD) Main Economics Indicators. Note that for real GDP per capita and import share of GDP data, only annual data is available.
Table 1: Calibrated Parameters
Parameter Description Value
Preferences
ϕ! Utility weights for real money balances 0.002
ϕ! Utility weights for leisure 11.2
β Discount factor 1.0059
η Inverse Frisch elasticity of hours worked 2
Production
λ! Mean productivity growth rate 1.016
ϕ Cost share of intermediate inputs 0.5
θ! Elasticity of substitution between differentiated goods 6
Ω! Price adjustment costs 61
International trade
α share of domestic inputs for import sector 0.7023
θ Export demand elasticity 1.5
Portfolio adjustments
Ω! Size of the portfolio adjustment costs 0.25
ψ Steady state portfolio share of domestic bonds in the total value of private bond holdings
0.82
Monetary Policy
λ! Mean depreciation rate 1.0085
Shocks and Persistence
ρ! Persistence of foreign interest rate shock 0.97
σ! Standard deviation of foreign interest rate shock 0.004
Technology growth is set to λ!= 1.016, so that real GDP per capita annual growth in steady state is around 6.5%. Note that this value is obviously much lower than the past 10-year average annual growth for China, which is amounted to over 9%. Yet, the fast growing pace reflects more likely that China was in its upswing of the business cycle instead of on a steady-state track. This is verified by the slow down of the economy in the recent one or two years. Thus, 6.5% is assumed to be a reasonable value, which is also close to the China’s parliament approved full-year growth target of 6.5-7% for 2016. With a steady-state technology growth rate included, the steady state real interest rate is pinned down by !!!. Targeting annual real interest rate of 4% in steady state, this implies β = 1.0059. The nominal money growth rate λ! is related to the steady
21 state inflation, π =!!
!!. λ! is set to λ! = 1.03, which implies an annul money growth of 12%,
close to the sample average for M2 of 16%, and is almost identical to the past 4-year sample average. Thus, the steady state inflation is determined: π = 1.0138, which implies annual inflation rate of 5.52%. With 4% real interest rate, the steady state nominal interest rate is also pinned down: R = 1.024. The steady state foreign inflation is calibrated at 1.0052, equivalent to 2.08% annual inflation rate, which is the sample mean of US inflation rate from 2005 to 2015. From the relationship !
!∗ =
!
!∗ in steady state, foreign interest rate in steady state is therefore
R∗ = 1.015. The implied steady state depreciation rate is λ
!= !!∗= 1.0085, which is identical to
around 3.4% annual depreciation rate – close to the depreciation rate of CNY against USD in 2015.
For the rest household preference parameters, the inverse Frisch elasticity of labor supply is set to η = 2 following Chang, Liu, & Spiegel (2015). The utility weight for leisure ϕ! is set to
ϕ! = 11.2, targeting the steady state fraction of labor hours of 40% of total time endowment. The utility weight for money is assumed to be sufficient low to ensure that the steady state domestic bond holding is positive, taking the value of 0.002.
For domestic production parameter, the cost share of intermediate inputs in production is set to ϕ = 0.5 in order to match the unit labor cost of 0.45 in steady state. While there is no data available for the unit labor cost for China, the average of unit labor cost for OECD countries is very robust for the past 20 years, which stays around 0.5 to 0.6. Given relative low costs of labor in China, a unit labor cost of 0.45 is considered reasonable. The elasticity of substitution between differentiated goods is set to θ!= 6, which implies a steady state price markup of 20%. I further calibrate the price adjustment cost as Ω!= 61, so that the corresponding average duration of
price contracts is 4 quarters. Specifically, in an economy with Calvo-pricing, the slope of the Phillips curve is given by !!!!! !!!!
!! , where ξ! is the probablity that a firm cannot repotimize
its price. Targeting average duration of price contracts for 4 quarters means that ξ! = 0.75, and that the slope of the Phillips curve is 0.0819. Equating the slope of the Phillips curve (!!!!
!! )
implied by my model with 0.0819 gives the value of Ω!.
For the parameters in the external sector, the share of domestic intermediate input is set to α = 0.7023, so that the import-to-GDP ratio is 24.36%, which is the mean value for China’s
22 import-to-GDP ratio from 2005 to 2015. The export demand elasticity θ is set to θ = 1.5 following Chang, Liu, & Spiegel (2015).
To calibrate the steady state portfolio share of domestic bonds, I use outstanding domestic government bond data to estimate domestic bond holding from the private household. For private holding of foreign bond, I use the data on asset value of portfolio investment in terms of debt securities, which is taken from the financial account portfolio investment of China’s balance of payment. Specifically, on the asset side of financial account, net acquisition of foreign asset can take the form of direct investment, portfolio investment, and other investment. For the purpose of the study, portfolio investment is considered to be the most relevant for private portfolio holdings. And since that only bond market is considered in the model, portfolio investment in terms of forms other than debt securities are not considered. The data on total asset of debt securities portfolio investment is only available quarterly since 2010. Before 2010, there is only quarterly data on the portfolio outflow. Thus, I construct the data before 2010 by adding the previous year-end asset by the flow value in the subsequent quarters. To this end, I construct the quarterly data from 2005 to 2015 on the portfolio share of private holding of foreign bond by combing the data on private holding of foreign bond and domestic government bond outstanding (both data are in USD). The steady state portfolio share of domestic bond is therefore calibrated as the mean ψ = 0.82, which lies in the range in the literature from, for example, Coeurdacier & Rey (2013), who found that average bond home bias worldwide in 2008 is equal to 0.75.
Portfolio adjustment costs is calibrated by regressing the following equation:
R!!!− R∗!!!− λ!,!= a + bψ!!!, (56)
where the China and US short term interest rate are used for R! and R!∗ respectively, and the
depreciation rate of CNY with respect to USD is used for λ!,!. The estimated portfolio share for China’s private sector described above is used for ψ!. The regression equation corresponds to the log-linearized version of equation (38):
R!− R∗!− E!λ!,!!! = Ω!ψψ!. (57)
The estimated value for b is 0.2044,which is significantly different from zero with a t-value of 9.7. The result indicates that the UIP condition does not hold in China. And the positive sign for the portfolio share shows that the relative demand for domestic bond is negatively related to relative price of domestic bonds, which is consistent with the intuition. With b = Ω!ψ, the
23 implied value for portfolio adjustment costs is Ω! = 0.25, which is slightly higher than the one estimated by Chang, Liu, & Spiegel (2015) for other emerging market economies.
Finally, for the shock parameters, the autoregressive coefficients for the foreign interest rate are ρ! = 0.97. The standard deviations for the shock is σ!= 0.004.
6. Monetary policy and hypothesis
6.1. The trade-offs of using capital control under pegged exchange rate
In the model, the use of capital controls has two effects on the real economy. The first one is the level effect. The increase in foreign interest rate will inevitably cause capital outflows as shown by the generalized UIP condition, which will be accompanied by increasing net export and real exchange rate depreciation implied by the balance of payments. Without adjustment in exchange rate, domestic inflation has to go down according to the real exchange rate determination equation. Since there is nominal rigidity in the economy, decrease in inflation is associated with lower real marginal cost according to the intermediate firm’s price setting decision equation (Philips curve). Decline in real marginal cost accompanied by depreciation implies that real wage drops according to equation (44). Consequently intermediate goods firms demand more labor compared to importing firm’s output as a result of the substitution effect. Diminishing demand for the importing firm’s output decreases importing firm’s demand for domestic product inputs, which results in another round of decreasing in domestic production and ultimately real GDP. It can be conjectured that the larger the portfolio adjustment costs, the smaller the capital outflows, and the smaller the negative effects of capital outflows on domestic economy caused by the positive interest rate shock.
Another effect is inter-temporal substitution effect, which can be seen from the log-linearized Euler equation:
c!− E!c!!!+ R!− E!π!!! = Ω!ψψ!(1 − ψ −!!! ), (58) and the log-linearized first order condition equation with respect to private holdings of foreign bond:
c!− E!c!!!+ R∗!− E!π!!!+ E!λ!,!!! = −Ω!ψψ!(ψ +!!! ). (59)
A shock such as increase in foreign interest rate will cause an inter-temporal shift of domestic consumption. Yet, there are two opposite effects at work, as already mentioned before, leaving
24 the net inter-temporal effects somewhat unclear. Specifically, declining in the holding of domestic bond (compared to steady state) reduces the cost of portfolio adjustment, whereas increasing in the holding of foreign bond increases the cost of portfolio adjustment. This can be shown from the fact that when 𝜓!< 0, the right hand side of the equation (58) is negative
whereas the right hand side of the equation (59) is positive.
To show what does the net effect depend on, I combine the two equations and get a reduced form of Euler equation:
c!− E!c!!!− E!π!!!= −R!ψ + R∗!+ E
!λ!,!!! ψ − 1 . (60)
Letting the left hand side positive is equivalent to have
R! < (R∗!+ E!λ!,!!!) 1 −!! . (61) Ignoring the change in 𝐸!𝜋!!! and let 𝐸!𝜆!,!!! = 0 for pegged exchange rate, one can see that as long as monetary policy responds aggressively to the positive foreign interest rate shock, there will be net shift from future consumption to present, which can dampen the negative effects from capital outflows on the real economy mentioned before. In fact, it is exactly the existence of capital controls that gives monetary authority the independence to do this. Moreover, the larger the capital controls i.e. higher portfolio adjustment costs, the larger the room the monetary authority has to change the interest rate, and the higher the benefits of applying capital controls.
The above analysis leads to the following testable hypothesis:
Hypothesis 1: Given fixed exchange rate, higher capital controls can have beneficial effects on reducing the negative impacts of capital outflows caused by the positive foreign interest rate shock both through the level and substitution effects.
6.2. The role of foreign exchange rate regime
The previous analyses apply to the case of pegged exchange rate regime, it can be easily shown that, however, the manage exchange rate regime can potentially complement the capital control policy. If increase in foreign interest rate is accompanied by nominal exchange rate appreciation, then less capital outflows will occur, so does the negative level effects of positive foreign interest rate shock on the domestic economy.While China is still virtually maintained pegged exchange rate regime with frequent revising, it is hard to calibrate a managed exchange rate policy to represent China since it is counterfactual. Nonetheless, Figure 1 shows that the exchange rate
25 change in China tends to move to the opposite of domestic inflation rate. That says, it seems that the exchange rate management is more inward looking: When inflation is low, it tends to go up (depreciation) to restore the economy. A simple regression of nominal exchange rate change on inflation rate using data from 2005 to 2015 also verifies this observation, and the coefficient has a negative value of -0.16 and is significantly different from zero with a t-value of -4.94. Yet, based on the analysis before, I argue that the current “managed exchange rate” is not suitable for an external shock with foreign interest rate increasing since that exchange rate should appreciate instead of depreciate as implied by the empirical evidence. I thereby make my second and third hypothesis:
Hypothesis 2: Given capital controls, managed exchange rate is a superior choice over fixed exchange rate regime in terms of shielding the economy from external shocks.
Hypothesis 3: A managed exchange rate that responds to external environment does a better job than the one responds to internal macroeconomic in shielding the economy from external shocks.
Figure 1. Dynamics of inflation rate and changes in nominal exchange rate (CNY/USD), 2005Q1-2015Q4
0.92 0.94 0.96 0.98 1 1.02 1.04 1.06 -10.0 -8.0 -6.0 -4.0 -2.0 0.0 2.0 4.0 6.0 8.0 10.0 12.0 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 n omin al exchan ge rate depreciation in 1lation rate , % in0lation rate, % depreciation rate
26 7. Results
7.1. Role of capital controls
In the benchmark scenario, exchange rate is considered to be fixed λ!,! = 0, domestic monetary authority follows a Taylor type rule R! = 1.5π!+ 0.5GDP!4, level of capital controls is represented by Ω! = 0.25.
Figure 2. Impulse responses to an increase in the foreign interest rate shock under baseline scenario5
Figure 2 shows the impulse response dynamic of the benchmark scenario. The 1 standard deviation positive foreign interest rate shock triggers capital outflows resulting from the portfolio rebalancing of private household. Consequently, real exchange rate depreciates. The real exchange rate depreciation has two opposite effects. On the one hand, domestic product becomes relatively cheaper than the foreign product. Thus, exports increase and the imported products share in the importing sector decreases. In combination, the net exports increases, which has
4
The use of Taylor rule here can be justified by the fact that China has been making efforts to strength the effectiveness in interest rate channel as explained in section 3.
5
All dynamic impulse responses figures in this paper are in unit of log deviation from steady state
0 10 20 30 ×10-4 -14 -12 -10 -8 -6 -4 Real GDP 0 10 20 30 ×10-3 -4 -3.5 -3 -2.5 -2 -1.5 Consumption 0 10 20 30 ×10-3 -3 -2.5 -2 -1.5 -1 -0.5
0 Nominal interest rate
0 10 20 30 ×10-4 -15 -10 -5 0 5 Inflation 0 10 20 30 -0.035 -0.03 -0.025 -0.02 -0.015 -0.01 -0.005 Portfolio share 0 10 20 30 ×10-3 -5 -4.5 -4 -3.5 -3 -2.5 -2 -1.5 Intermediate output 0 10 20 30 ×10-3 -7 -6.5 -6 -5.5 -5 -4.5 -4 -3.5 Import 0 10 20 30 ×10-3 1 2 3 4
5 Real exchange rate
0 10 20 30 0 0.05 0.1 0.15 0.2 0.25 Current account
27 expansionary effects on the real economy. One the other hand, real exchange rate depreciation, together with the fixed exchange rate, implies that inflation should goes down. Because of the existence of nominal rigidity, the change in inflation rate has negative effects on output and consumption. In my calibration, the latter effect dominants the former, leaving the net effect of positive foreign interest rate shock on the economy to be detrimental to the domestic real economy, reflecting by decline in real GDP, which is consistent with intuition. As explained before, the existence of portfolio adjustment costs allow domestic monetary policy to respond to the external shocks, which is reflected in the expansionary monetary policy immediately after shocks. Moreover, portfolio adjustment costs affect household inter-temporal decision in consumption. This can be seen from the dynamics of the consumption that it first declines and then reverts back to the steady state. This implies that at the time of shock, the net increase in portfolio adjustment costs makes consumers to shift future consumption to present, which, to some extent, offset the immediate detrimental effects of the foreign interest rate shock.
To test hypothesis 1, I vary the portfolio adjustment costs parameter to investigate the effects of capital controls. Specifically, I conduct simulation with portfolio adjustment costs at 0.1 and 0.4 respectively. The main comparison variables are change in real GDP and inflation. The result is presented in Figure 2. The immediate response of variables is as predicted, thought not significant. When capital control is loosened, i.e. Ω! = 0.1, there are larger capital outflows
compared to the baseline scenario as can be seen from the larger drop in portfolio share. Large capital outflows cause relatively more real exchange rate depreciation and thus more decrease in inflation. Yet it shall be noticed that GDP actually declines less than the baseline scenario, reflecting the fact that the benefits of more real exchange rate depreciation dominates the costs. Nonetheless, the change in the level effects of capital outflows is not very significant. This may highlight the role of government balance sheet. As the demand for money and domestic government bond shrinks dramatically, the government has to sell large amount of its foreign bond holding. This partially offsets the effects of private capital outflows, leaving the effects of change in private capital controls not that significant. While the level effect of varying capital control is very small, the reduction in substitution effect caused by declining capital control is evident. The inter-temporal shift in consumption is also much smaller due to the smaller portfolio adjustment costs.
28 The immediate effect of higher capital controls, i.e. Ω! = 0.4 is in general the opposite compared to the scenario with loosened capital controls, besides that the change is relatively small, reflecting the asymmetric effects of varying levels of capital controls.
Figure 3. Impulse responses to an increase in the foreign interest rate shock under different level of capital controls
It is worth noticing that the longer time effects of having higher capital controls are more damaging, as can be seen that the scenario with lower capital control returns to the steady state much quicker than the baseline and higher capital control scenarios. This reflects the flip side of capital controls. While capital controls limit the immediate shift in private portfolio, it also hampers the adjustment in the portfolio in the future. In contrast, the scenario with much lower portfolio adjustment costs allow the portfolio to return to the steady date much faster. Moreover, while the portfolio adjustment costs allow the shift of consumption from future to present, it also means that some of the shocks are extended to the future, which also slow down the process of returning to steady state. Although I do not calculate the welfare effects directly, with my calibration of subjective discount factor higher than 1, I tend to believe that less capital controls is a more welfare enhancing option.
0 10 20 30 ×10-4 -14 -12 -10 -8 -6 -4 -2 Real GDP 0 10 20 30 ×10-3 -4 -3.5 -3 -2.5 -2 -1.5 -1 Consumption 0 10 20 30 ×10-3 -3 -2.5 -2 -1.5 -1 -0.5 0
0.5 Nominal interest rate
0 10 20 30 ×10-4 -15 -10 -5 0 5 Inflation 0 10 20 30 -0.1 -0.08 -0.06 -0.04 -0.02 0 Portfolio share 0 10 20 30 ×10-3 -6 -5 -4 -3 -2 -1 0 Intermediate output 0 10 20 30 ×10-3 -7 -6 -5 -4 -3 -2 Import 0 10 20 30 ×10-3 1 2 3 4
5 Real exchange rate
0 10 20 30 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 Current account
29 In conclusion, hypothesis 1 is verified that indeed when capital controls are higher, both the level effects and substitution effects contribute to smaller movements in real GDP and inflation. However, the immediate benefits from higher capital control are very limited as reflected by little change in movements in real GDP and inflation when capital control increases. What is more, the long run effects of having higher capital controls seem to be more damaging both because of level and substitution effects. With a subjective discount factor higher than 1, it is more likely that the damaging long term effects will dominate the short term benefits of having higher capital controls.
7.2. Managed exchange rate
To test hypothesis 2 and 3, I compare two alternative managed exchange rate regimes with the benchmark scenario keeping the other conditions unchanged. For simplicity, I assume exchange rate only responds to inflation rate and assume the following two opposite counterfactual exchange rate rules: λ!,!= 0.5π! (rule 1) and λ!,!= −0.5π! (rule 2). The first one is hypothesized to be better than the second one, as it responds in the right direction after the shock, i.e. appreciation, which is seen as an external looking rule. The second one is more in line with the current relation between depreciation rate and inflation rate in China. The parameter here is just hypothetical, in fact, changing the parameter does not vary the results presented below.
Figure 4 shows the impulse response of the two alternative exchange rate rules and the benchmark scenario. Compared the fixed exchange rate regime with managed exchanged rate rule 1, appreciation in nominal exchange rate helps shield domestic economy from external interest rate shock on the real economy as expected. Specifically, the reduction in GDP declines by a significant amount. Yet, inflation declines more, which is not surprising since that nominal exchange rate appreciates. On the contrary, managed exchange rate rule 2 has the opposite effects: real exchange rate depreciates more leading to less decrease in inflation with the compensation of more decreases in GDP. Since that the benefit of using rule 2, which induces less variation in inflation, is much smaller than its cost, which induces more variation in real GDP, rule 1 seems to be a better choice than rule 2.
30
Figure 4 Impulse responses to an increase in the foreign interest rate shock under different exchange rate regimes
Thus, for hypothesis 2, it is indeed that with exchange rate rule 1, variation in GDP is much smaller than the baseline scenario with fixed exchange rate though variation in inflation is slightly more. Given the benefit of having much smaller real GDP variation with exchange rule 1, one can fairly say that managed exchange rule 1 is better than the fixed exchange rate. However, using managed exchange rule 2 is more harmful with more declines in real GDP but much less decline in inflation than both managed exchange rule 1 and the baseline scenario. This is consistent with hypothesis 3 that exchange rate rule that responds to the external shock does a better job in maintaining stability than the exchange rate rule that responds to the domestic economy. And the result further asserts that with an internal looking managed exchange rate rule 2, the result is even worse than the pegged exchange rate regime. This finding is consistent with the literature in optimal monetary stabilization policy in open economies with incomplete market
included mentioned in the literature review. For example, as explained in the literature review, Corsetti, Dedola, & Leduc (2010) also find that when there is incomplete asset market, internalize international monetary spillovers requires the targeting rules to stress external
0 10 20 30 ×10-3 -1.6 -1.4 -1.2 -1 -0.8 -0.6 -0.4 Real GDP 0 10 20 30 ×10-3 -4 -3.5 -3 -2.5 -2 -1.5 Consumption 0 10 20 30 ×10-3 -3 -2.5 -2 -1.5 -1 -0.5
0 Nominal interest rate
0 10 20 30 ×10-4 -15 -10 -5 0 5 Inflation 0 10 20 30 -0.035 -0.03 -0.025 -0.02 -0.015 -0.01 -0.005 Portfolio share 0 10 20 30 ×10-3 -6 -5 -4 -3 -2 -1 Intermediate output 0 10 20 30 ×10-3 -7 -6 -5 -4 -3 Import 0 10 20 30 ×10-3 0 1 2 3 4 5
6 Real exchange rate
0 10 20 30 0 0.05 0.1 0.15 0.2 0.25 0.3 Current account
31
conditions such as misalignments in the terms of trade and real exchange rates, and cross-country demand imbalances.
8. Conclusion
The present paper uses a small open economy New Keynesian model with four characteristics of China’s policy framework included in the benchmark model. Specifically, it considers (1) a pegged exchange rate regime; (2) central bank’s flow-of-fund constraint to reflect the role of government intervention in the foreign exchange rate market; (3) an incomplete asset market where foreign investors are not allowed to hold domestic bond and domestic investors have portfolio adjustment costs when adjusting their portfolio; and (4) price rigidities and imperfect competition so that monetary disturbances have real effects. The paper thus examines the impluse response of a positive foreign interest rate shock on the domestic economy in terms of stability in real GDP and inflation.
By adjusting portfolio adjustment costs, I find that when capital controls are higher, both the level effects and substitution effects contribute to smaller movements in real GDP and inflation. However, the immediate benefits from higher capital control are very limited as reflected by little change in movements in real GDP and inflation when capital control increases. Yet, the long run effects of having higher capital controls seem to be more damaging. With a subjective discount factor higher than 1, it is more likely that the damaging long term effects will dominate the short term benefits of having higher capital controls.
I further compare two alternative exchange rate rules namely the managed exchange rule that responds to the external shocks, and the managed exchange rule that is inward looking. I find that the former is better than fixed exchange rate regime, and complement capital controls in stabilizing the economy, particularly real GDP, while the latter performs the opposite.
The findings in the paper suggest two policy implications. First, the benefit of further rising capital controls is very limited while lower the capital controls seems to be significant, especially in the long run. Thus, it is suggested that the China’s government should continually easing its capital controls. Second, instead of using a pegged exchange rate regime, a managed exchange rate that is outward looking is a better instrument in shielding domestic economy from external shocks. Thus, the China’s government should make use of exchange rate as an instrument in shielding domestic economy from foreign interest rate increasing shocks.
