# Brexit associated sharp depreciation and implications for UK’s inflation and balance of payments

Muhammad Ali Nasir (Department of Economics, Leeds Beckett University, Leeds, UK)
Justine Simpson (Department of Finance and Accounting, Leeds Beckett University, Leeds, UK)

ISSN: 0144-3585

Publication date: 14 May 2018

## Abstract

### Purpose

The purpose of this paper is to analyse the implications of exchange rate depreciation for inflation targeting and trade balance of UK in the context of the Brexit epoch.

### Design/methodology/approach

The study employed a time-varying structural vector auto-regression (TVSVAR) model framework in which the sources of time variation were both the coefficients and variance-covariance matrix of the innovations on the data from January 1989 to September 2016.

### Findings

The findings suggest that the depreciation of the Stirling has significant effects on inflation and trade balance in UK in context of Brexit epoch. It also showed that such a depreciation can be helpful in the improvement of external balance as well as steering the inflation to its statutory target. Despite, the inflation targeting, there is strong evidence of a pass-through.

### Research limitations/implications

Research has profound implications in terms of the sharp depreciation of GBP associated with the Brexit outcome. The study is very topical and could be very interesting to the readership of JES as well as wider audience. The study has limitations in a context that the significance of the results and association of the under analysis entities is contingent on the future trade relationships and Channel between UK and EU. Therefore, although there is a lot of uncertainty about the future of Britain trade relationships, this study provides guidance on the importance of exchange rate channel if the similar trade arrangements prevails in the post-Brexit era.

### Practical implications

The research has profound practical implications, using a TVSVAR model in which the relationship among the entities varies over time; it has shown the importance of exchange rate in terms of external balance and inflation targeting. Hence, it has appeal for the practitioners as well as academics.

### Social implications

The research has great social implications. The Brexit is the biggest political and economic event of this era for UK and EU. There are big questions about the relationship between UK and EU in the post-Brexit epoch as well as questions about the future of the European integration. In this context, this study has shown that how the exchange rate could play an important role for the UK economy when its contemporary trade channels prevail. Concomitantly, it has social implications particularly for the European society.

### Originality/value

The research is an original piece of work. It has contributed to the debate on the exchange rate deprecation, external balance and inflation targeting in context of the Brexit associated sharp depreciation of Stirling. It has used a framework, i.e. TVSVAR, which also have unique features in terms of testing the associations among under analysis entities against time.

## Keywords

#### Citation

Nasir, M. and Simpson, J. (2018), "Brexit associated sharp depreciation and implications for UK’s inflation and balance of payments", Journal of Economic Studies, Vol. 45 No. 2, pp. 231-246. https://doi.org/10.1108/JES-02-2017-0051

### Publisher

:

Emerald Publishing Limited

## 1. Introduction

Brexit is one of the most significant events of recent political and economic history with implications for the European Union and the British economy in particular[1]. The outcome of the referendum was to leave the EU, which led to a wave of uncertainty in the economic and financial world and among other reactions; the British Pound incurred one of the sharpest deprecations in its history. Against the US $, the GBP £ depreciated to its lowest level in 31 years and moved to £1=$1.21. Concomitantly, among others factors, this sharp depreciation could have profound implications for the British economy and the financial sector which by any measure are nontrivial in terms of their size and significance. The British economy is the fifth largest economy in the world and its financial sector is top of the list in the Global Financial Centre Index (GFCI, 2016).

A bird’s eye view of the macroeconomic outlook of the British economy in recent history makes two factors prominent which are also the focus of this study. These are the low level of inflation and huge current account deficit. Inflation has been persistently low for a number of years but particularly in the last two years, it has been below the 2 per cent rate which is the statutory target given to the Bank of England (Figure 1).

In the external balance, Britain has also been running one of the largest trade deficits where its trade deficit reached the highest in history. The UK’s current account deficit was over £25 billion in the third quarter of 2016, equating to 5.2 per cent of gross domestic product (GDP) at current market prices, the largest proportion since quarterly records began in 1955, up from 6 per cent in the fourth quarter of 2013 (Office for National Statistics, 2016.

Despite the bleak scenario presented above of lower than target inflation and a high current account deficit, the depreciation of the GBP (£) has a silver lining. Theoretically, the depreciation of a currency can have a positive impact on inflation (pass-through) as well as the trade balance. On this notion, one can go back as far as Hume’s (1742) price-specie flow mechanism argument[2]. However, the net benefit of depreciation can only be positive if the elasticities of exports and imports sum up to a value greater than unity, i.e. the Marshall-Learner condition (see Bahmani-Oskooee and Ratha, 2004 for interesting insight). The existing evidence on the subject suggests that there can be considerable improvements made in the external balance of both developed and developing countries (see Bahmani-Oskooee, 1991; Bahmani-Oskooee and Ratha, 2004; Bahmani-Oskooee and Hegerty, 2010; Bahmani-Oskooee et al., 2013, 2016; Yildirim and Ivrendi, 2016, discussed further in the next section). However, there are also some studies which failed to find significant evidence of exchange rate movements on the balance of payment, for instance, Rose and Yellen (1989) and/or Rose (1991). This study adds to the debate by analysing the implications of the exchange rate for the UK balance of trade which is running the largest deficit in the context of Brexit associated sharp depreciation. Second, we are also investigating the implication of exchange rate depreciation for inflation which has been below even the modest target of 2 per cent, posing a risk of deflation. On this aspect, there are some remarkable studies, for instance, Bahmani-Oskooee (1991), Wimanda (2014) and Yildirim and Ivrendi (2016) which reported a positive impact of exchange rate depreciation on inflation (detailed discussion in next section). An important contribution of this study is endeavouring to make to the debate on the exchange rate and inflation nexus is by considering the time-variant aspect of this relationship. On this issue, Forbes (2015) argued that “One assumption that is traditionally made in many institutions when analysing pass-through is that it is fairly stable over time – at least within a given country. This assumption has continued to be used in most analyses of pass-through, despite it being challenged in a number of studies over the years – including at the Bank of England” (Forbes, p. 14). Concomitantly, we are considering the time-varying association between the exchange rate and inflation, nevertheless, we are also including the balance of trade into the analysis as the relationship between the exchange rate and trade balance may also vary over time. The underlying logic of that could be the increasing cross-border trade and economic integration which varies over time. Lastly, the UK is an economy where inflation has been explicitly targeted for the last 25 years[3], on this aspect, there is a notion that the inflation targeting could diminish the impact of the exchange rate on inflation (see Edwards, 2007). In this study, we will put this notion to the empirical testing and see if the importance of exchange rates for inflation has diminished due to inflation targeting.

Putting all the aspects together, specifically, in this study, we analysed the impact of real effective exchange rates (REERs) for inflation and the balance of payments of the British economy in a time-varying setting. In order to do so, we employed a time-varying structural vector auto-regression model (hereafter TVSVAR) in which the sources of time variation were both the coefficients and variance-covariance matrix of the innovations on the data from January 1989 to September 2016. The key findings suggest that the sharp depreciation of Stirling is associated with the surge in inflation, which could help to bring inflation back to its target in a regime where it has been persistently low. Despite, the inflation targeting, there is strong evidence of pass-through. The depreciation of Stirling also leads to considerable gains in the external balance as it can also help to bring considerable improvements in the balance of payments. There was also strong evidence from the J-curve. Looking forward, any fruits which the exchange rate adjustments may bring in the future will be contingent on the outcome of the Brexit negotiations and future trading relationships of the UK.

The paper proceeds as follows, Section 2 provides a discussion on the nexus between exchange rates, inflation and trade balance, Section 3 briefly describes the layout of the empirical framework, Section 4 presents the findings and Section 5 draws conclusions and discuss the policy implications.

## 2. Depreciation, prices and balance of trade

On the impact of exchange rates on inflation, Bahmani-Oskooee (1991) established that the current floating exchange rate system, exchange rate variability is another factor contributing to inflation variability. A comprehensive empirical analysis was conducted in their study using cross-country data from 20 developed and 76 less developed countries. In later studies, Ali and Anwar (2016) reported that significant effects on prices, especially input imports. There are a number of countries which explicitly target inflation in specific to the UK at 2 per cent of Consumer Price Index (CPI) annually. The significance of an inflation target is that unlike monetary targeting, it does not depend on the stability of the demand for money and, unlike foreign exchange targeting, does not require changes in interest rates, direct foreign exchange intervention and the loss of independent monetary policy (see Mishkin, 1998; Canarella and Miller, 2016). On the aspect of inflation targeting, a study by Gerlach (1994) argued that there is a symmetry implying the importance of exchange rates in this regard. Although, Pourroy (2012) argued that a managed exchange rate environment delivers a stronger nominal anchor to inflation shock and concomitantly helps in the inflation targeting, in specific to the UK, the exchange rate is not managed but inflation has a statutory target, hence depreciation has implications for price stability. On the idea of a flexible exchange rate, Ghosh et al. (2013) showed that flexible exchange rates could be more effective in the external balance adjustment[7]. Perhaps, on this aspect in the UK, we are not faced with the “fear of floating” highlighted by Calvo and Reinhart (2000), as there is a free float. Interestingly, although Calvo and Reinhart (2000) argued that the pass-through of inflation from exchange rate swings is larger in the emerging markets as compared to developed countries, there was a strong opposition to exchange rate pegging put forward by Mishkin (1998). The question of managed or float is beyond the scope of this study, yet in the given exchange rate regime and inflation targeting regime, the question is what implications depreciation can have for inflation in the UK. A study by Edwards (2007) while investigating the relationship between inflation and exchange rate targeting argued that the countries that have adopted inflation targeting have experienced a decline in the pass-through from exchange rate changes to inflation. Moreover, the adoption of inflation targeting monetary policy procedures has not resulted in an increase in the volatility of exchange rates. Similarly, Gagnon and Ihrig (2004), on developed countries, argued that the pass-through in developed economies has declined due to inflation targeting. The subject study will give us further insight into what implications it could have for the UK where we have an explicit and statutory inflation-targeting framework. Does the exchange rate depreciation still have some implications for inflation and specifically in the context of the Brexit saga? In particular, we are employing a time-varying framework that will cater for the time variation aspect, covering the period of the UK moving to inflation targeting as well as the events of Brexit. On the choice of framework, Balcilar et al. (2016) urged the use of the time-varying parameter model while analysing the impact of supply and demand shocks to the exchange rate, although in our case the line of reasoning and direction of causality is rather from exchange rates to inflation and trade balance, yet the time-varying framework is appropriate to give us an insight into the under analysis axioms which can hold against time variation. The rationale of considering the time-contingent association of economic entities is manifested in the reasoning of Keynes (1938) as:

[E]conomics is a science of thinking in terms of models joined to the art of choosing models which are relevant to the contemporary world. It is compelled to be this, because, unlike the typical natural science, the material to which it is applied is, in too many respects, not homogeneous through time. The object of a model is to segregate the semi-permanent or relatively constant factors from those which are transitory or fluctuating so as to develop a logical way of thinking about the latter, and of understanding the time sequences to which they give rise in particular cases.

Concomitantly, acknowledging the weaknesses of axioms in the face of tides of time, we are employing a time-variant framework. The next section will elaborate it further.

## 3. Methodology

In order to analyse the impact of REERs on the balance of payments and inflation in the UK, we are considering a time-varying structural vector autoregressive (TVSVAR) model which is based on the seminal work by Primiceri[8] (2005). The beauty of this framework is that both the coefficients and the entire variance-covariance matrix of the shocks are allowed to vary over time. This is crucial if the objective is distinguishing between changes in the typical size of the exogenous innovations and changes in the transmission mechanism of exchange rate shocks (Primiceri, 2005). It is suitable to capture changes in private sector behaviour, where aggregation among agents usually plays the role of smoothing most of the changes. Specifically, the existence of any type of learning dynamics in terms of association among under analysis entities (exchange rates, the balance of trade and inflation) seems to favour a model with smooth and continuous drifting coefficients and heteroscedasticity innovations.

### 3.1 The model

The TVSVAR model employed in this study has properties of both time-varying coefficients and a time-varying variance-covariance matrix of the additive innovations. It is useful in the sense that drifting coefficients capture possible nonlinearities or temporal variation in the lag structure of the model. Nevertheless, the multivariate stochastic volatility captures the possible heteroscedasticity of the shocks and nonlinearities in the simultaneous relations among the variables of the model. Given that the time variation is allowed both in the coefficients and the variance-covariance matrix, leaves it up to the data to determine whether the time variation of the linear structure derives from changes in the size of the shocks (impulse) or from changes in the propagation mechanism (response). A point to be noted here is that the TVSVAR framework admits many types of shocks (for details see Cogley and Sargent, 2003; Primiceri, 2005). Let us consider the following model:

(1) y t = C t + B 1 , t y t 1 + + B k , t k + u t t = 1 , , T
where yt is an n × 1 vector of observed endogenous variables; specific to this study, these will be exchange rates, inflation and the current account balance; Ct is an n × 1 vector of time-varying coefficients that multiply constant terms; Bi,t,i = 1, …, k are the n × n matrices of time-varying coefficients; ut are heteroscedastic unobservable shocks with variance-covariance matrix Ωt defined by:
(2) A t Ω t A t = t t ,
where At is the lower triangular matrix:
A t = [ 10 0 a 10 1 0 a n1 , t a n n 1 , t 1 ]

And t is the diagonal matrix:

t = [ σ 1 , t 0 00 σ 2 , t 00 0 σ n , t ]

Thus, it follows that:

(3) y t = C t + B 1 , t y t 1 + + B k , t k + A t 1 t ε t
V ( ε t ) = I n

Stacking in a vector Bt all the right hand side coefficients in Equation (3) can be rewritten as:

(4) y t = X t B t + A t 1 t ε t ,
X t = I n [ 1 , y ( t 1 ) , ... , y ( t k ) ]
where the symbol ⊗ denotes the Kronecker product. It is common to decompose the variance-covariance matrix as occasioned in Equation (4), this practice is more common in the studies focused on the problem of efficiently estimating covariance matrices (see Smith and Kohn, 2002; Primiceri, 2005). Similar decomposition is also evident by studies, e.g. Cogley (2003) and Cogley and Sargent (2003) employing time-varying VAR models, however, with a time invariant At matrix. As we have discussed earlier, it is vital to allow the matrix At to vary over time for a time TVSVAR framework. Keeping the At constant would imply that an innovation to the ith variable has a time invariant effect on the jth variable. This is definitely undesirable if we are aiming to model the time variation in a simultaneous equation model, particularly, where simultaneous interactions among variables are fundamental as in this study[9]. At this juncture, if we let the at be the vector of non-zero and non-one elements of the matrix At and σt be the vector of the diagonal element s of the matrix Σt, the dynamics of our model’s time-varying parameter can be specified in the following fashion:
(5) B t = B t 1 + v t ,
(6) a t = a t 1 + ζ t
(7) log σ t = log σ t 1 + η t
where the elements of the vector Bt are modelled as random walks, as well as the free elements of the matrix At[10]. Although the random walk process might be considered undesirable here due to the general perception that it hits any upper or lower bound with probability one. However, as long as the ((5)-(7)) are placed for a finite period of time this set of assumptions are innocuous. Nevertheless, the random walk assumption comes with the advantage of focusing on permanent shifts and reduced numbers of parameters (Primiceri, 2005). All the innovations in the model are assumed to be jointly normally distributed with the following assumptions on the variance-covariance matrix:
(8) V = V a r = ( [ ε t v t ζ t η t ] ) = [ I n 0000 Q0000 S0000 W ]
where In is an n × n-dimensional identity matrix Q, S and W are positive definite matrices[11]. The coefficients of the contemporaneous relations among variables are assumed to evolve independently in each equation, although it is not a crucial assumption, yet it simplifies the inference and increases the efficiency of the estimation algorithm.

### 3.2 Bayesian estimation

Estimation of model is carried out using the Bayesian approach which will be employed for evaluation of posterior distribution of the parameters of interest, in specific to our case BT, AT, T and the hyperparameters of the variance-covariance matrix V. The Bayesian approach is feasible in a scenario where the distinction between parameters and shocks is less clear while one is dealing with unobservable components[12]. Nevertheless, in this study, Gibbs sampling is employed for the posterior numerical evaluation of the parameters of interest. To start with, 48 observations (three years period) are chosen for a training sample. Thereafter, 10,000 iterations of Gibbs sampling will be chosen with a burn rate of 20 per cent, i.e. 2,000 iterations (so that the effect of initial values on the posterior inference is minimised.) Gibbs sampling is a particular variant of the Markov Chain Monte Carlo (MCMC) methods that consists of drawing from lower dimensional conditional posteriors as opposed to the high dimensional joint posterior of the whole parameter set. Considering that the MCMC is a smoothing method and hence delivers smoothed estimates, i.e. estimates of the parameters of interest based on the entire available set of data. Specifically, in a study like this where the objective is an investigation of the true evolution of the unobservable states over time, the smoothed estimates are more efficient and hence preferable (see Primiceri, 2005 for detailed discussion)[13]. Nevertheless, filtered estimates in the subject study are also not very appropriate because they would exhibit transient variation even in time invariant models as pointed out by Sims (2001).

### 3.3 Prior selection

The selection of prior distributions is based on their appropriateness and applicability. To start with, the assumption that the initial states for the coefficients, for the covariances, for the log volatilities and the hyperparameters are independent of each other is intuitive. The priors for the hyperparameters, Q, W and the blocks of S, are assumed to be distributed as independent inverse-Wishart. The priors for the initial states of the time-varying coefficients, simultaneous relations and log standard errors p(B0), p(a0) and p(logσ0), are assumed to be normally distributed. These assumptions together with (5), (6) and (7) imply normal priors on the entire sequences of the B’s, α’s and log σ’s (conditional on Q, W and S). The use of normal priors is fairly standard and also not being conjugate they have advantages in terms of tractability (see Smith and Kohn, 2002; Sims and Zha, 1998; Primiceri, 2005). As mentioned earlier, the MCMC algorithm is used to generate a sample from the joint posterior of BT, AT, ΣT and VT. Gibbs sampling is used in order to exploit the blocking structure of the unknowns and it is performed in four steps: drawing in turn time-varying coefficients BT, simultaneous relations AT, volatilities ΣT and hyperparameters V, conditional on the observed data and the rest of the parameters[14]. This empirical framework layout is applied to the data. The details of the data set are as follows.

### 3.4 Data set

The data set includes the series on the three variables of interest which are real exchange rate, inflation and trade balance, the details of each are as follows.

#### Inflation

For inflation, we used the CPI which is also the official target of the Bank of England monetary policy; hence, the most suitable proxy in this context.

#### Exchange rate

For exchange rate, the REER index is used, as it is the weighted average of the exchange rate of Sterling against the major trade partner’s currencies.

#### Balance of payment

The current account balance as a percentage of GDP was used as the proxy, the data available was quarterly observations which were converted into monthly observations by linear interpolation.

The data set includes the monthly observations from January 1989 to September 2016 (n=336). The data on inflation and the current account balance were obtained from the Office for National Statistics (ONS) while on the REER data were obtained from the Bank of England’s Bankstats.

## 4. Analysis and findings

The REERs are ordered last, it is due to the exogenous nature of exchange rate shocks. Moreover, the identification assumption also employs exchange rate shocks affects with lags. The simultaneous interaction between REER, CPI and BoP is arbitrarily modelled in a lower triangular form with the BoP first. It is not an identification condition but for the sake of normalisation, although the arbitrary normalisation may have the potential to make a difference; however, in this setting the ordering of the REER block did not affect the results. We chose 10,000 iterations of Gibbs sampling with a burn rate of 20 per cent, i.e. (2,000 iterations). The 48 observations (three years) period was chosen for a training sample. To start with, it is important to have a look at the time-varying standard deviation of the REER shocks. Figure 2 presents the plot of the posterior mean and the 16th and 84th percentiles of the time-varying standard deviation of the REER shocks. The percentiles correspond to the bound of a one standard deviation confidence interval.

Figure 2(a-c) gives some interesting insight into the behaviour of under analysis series, the balance of trade has been persistently volatile while inflation has been rather smooth in the pre- and post-GFC periods. It is also obvious that the GFC exchange rates did show higher variance which is understandable in terms of financial turmoil and high volatility. Thereafter we can witness a consistent oscillation and rather smoothing pattern which again increased in volatility as we approached the period of the Brexit saga and associated volatility.

The dynamics of the REER shocks and response of the trade balance are summarised in Figure 3(a-e). It presents the impulse responses of the trade balance to the REER shocks in four different dates of the under analysis sample. It also presents the pairwise difference between impulse responses in different dates with the 16th and 84th percentiles. The dates chosen for the comparison are August 2007, December 2013, June 2016 and December 2016. The choice of dates is due to the corresponding macro financial events and environment of prevailing periods. August 2008 is just before the GFC[15]. December 2013 corresponds to the period when the UK ran its largest trade deficit (6 per cent of GDP) since record began in 1956. June 2016 is the month of the Brexit referendum and December is the latest data available post-Brexit.

The results show the shock to the REER (appreciation), after an initial improvement there was a persistent worsening of the current account. This suggests a J-curve behaviour[16] (although in the opposite direction due to positive shocks) and implies that the depreciation of the currency in the real sense does bring international competitiveness for the UK economy and external balance. On this aspect, the findings are in line with studies for instance on USA and Indonesia by Bahmani-Oskooee and Harvey (2015) and among others, Bahmani-Oskooee and Ratha (2004), Bahmani-Oskooee and Hegerty (2010), Bahmani-Oskooee et al., (2013) and Yildirim and Ivrendi (2016). The findings are contrary to the Aristotelous (2001) analysis (limited to UK-US trade) and Rose and Yellen (1989) on the US data and Rose (1991) analysing the Post-Breton Woods period (1974-1986) of five OECD (UK, Canada, USA, Germany and Japan). Perhaps, this is prima facie evidence of the changes in the relationship between exchange rates and trade balance as economic integration and openness has increased over time. These findings also compliment and add to the work by Bahmani-Oskooee and Kovyryalova (2008) and Pattichis (2012) which were focusing on US-UK trade and a recent study by Bahmani-Oskooee et al. (2016) which was focusing on the UK and its trading partners (Germany, USA, Canada, Italy, Japan, Korea, Spain and Norway). There is strong evidence of J-curve behaviour which implies that the sharp depreciation can contribute to overcoming the adverse outlook of UK historical trade balance. Nonetheless, there were subtle differences in the periods pre- and post-global financial crises, which indicates that the relationship has grown stronger which could be associated with the increase in integration of economies as well as non-recessionary and crises periods in the global economy where the exchange rate depreciations can be more helpful. The pairwise difference (3B-3E) did not show much difference in the association in different periods which is an indication of the robustness of the relationship in the face of time variations.

The shock to the REER and the responses of inflation and the pairwise differences are presented in Figure 4(a-e).

The positive shock to the REER (appreciation) led to a persistent negative response from inflation. This implies a drop in the inflation rate due to the appreciation of GBP in real terms, putting this differently, depreciation leads to an increase in the inflation rate. The findings complement the studies, for instance, Bahmani-Oskooee (1991) which was focusing on less developed countries. Nonetheless, it adds to existing knowledge and provides an alternative perspective to the argument by Pourroy (2012) that a managed exchange rate environment delivers a stronger nominal anchor to inflation shock and concomitantly helps in inflation targeting, in specific to the UK the exchange rate is not managed but inflation has a statutory target, hence the depreciation has implications for price stability. Furthermore, the findings are also contrary to the Calvo and Reinhart (2000) argument that the pass-through of inflation from exchange rate swings is larger in the emerging markets as compared to developed countries. It also contradicts the notion by Gagnon and Ihrig (2004) and Edwards (2007) that the countries that have adopted inflation targeting have experienced a decline in the pass-through from exchange rate changes to inflation. Perhaps, it is prima facie evident that for the UK where we have an explicit and statutory inflation-targeting framework the deprecation has sustainable implications for inflation. Hence, to say least, the 25 years of inflation targeting in the UK may or may not have some success in taming inflation, however, our findings suggest that the nexus between inflation and exchange rate (pass-through) holds its grounds fairly strongly. Nonetheless, it is also obvious that in comparison to the pre-global financial crisis, the response of inflation was rather more pronounced which implies that exchange rate has become more effective for inflation dynamics. In fact, the recent and post-Brexit period showed a greater impact on inflation which indicates that the pass-through has been increasing over time. Furthermore, the increase in the pass-through over times also indicates the difference between the recessionary periods around the global financial crisis and the recent period when the economy has been growing. The pairwise difference did not show significant results implying the robustness of association in the face of time variation. It leads us to conclude in the next section.

## 5. Conclusion

On the basis of our empirical framework, we can hereby conclude that the exchange rate has significant implications for the UK’s external balance and price stability. Depreciation has a silver lining as it could lead to an increase in international competitiveness and rebalancing of the balance of trade deficit as well as a steering inflation to its target. The study also has wider economic implication beyond Brexit, as it implied that the J-curve behaviour of the trade balance holds. Nonetheless, it is also implied that the strategy of inflation targeting which has been adopted by some of the central banks including the Bank of England has not diminished the effects of depreciation on inflation (pass-through). However, we would acknowledge that these findings and implications should be taken with a pinch of salt as the future holds a lot of uncertainty around Brexit and future trading arrangements. Perhaps, these findings also have very important policy implication in terms of future trading relations between the UK and the European Union. A Soft-Brexit where the trade channels and trading arrangements are to a large extent analogous to the ones that prevailed during the EU membership may occasion the fruits of depreciation discussed in this treatise with higher probabilities of occurrence.

## Figures

#### Figure 1

CPI and balance of payment (C/A as % GDP)

#### Figure 2

Posterior mean, 16th and 84th percentiles of the standard deviation

#### Figure 3

(a) Real effective exchange rate shocks August 2007, December 2013, June and September 2016 (b-e) pairwise difference between the responses in corresponding periods with 16th and 84th percentiles

#### Figure 4

(a) Real effective exchange rate shocks August 2007, December 2013, June and September 2016 (b-e) pairwise difference between the responses in corresponding periods with 16th and 84th percentiles

## Notes

1.

In response to the referendum British voters, choose to leave the European Union.

2.

Hume (1742) while arguing in favour of free trade made the case against the mercantilist idea of having policy to run a favourable or positive trade balance. The price-specie flow mechanism states that countries with positive trade balances are effectively importing gold (money) in exchange for their exports while those with negative trade balances are exporting gold in exchange for imports. The increase in gold in countries with positive trade balances causes inflation, which makes prices rise and in turn makes imports more competitive. Conversely, the decrease in gold in countries with negative trade balances causes deflation, which makes prices fall and exports more competitive internationally. This causes the balance of trade to shift in both countries. Thus, Hume argued that a trade balance is relatively unimportant because it tends to balance itself out in the long term.

3.

The UK was the second country which started inflation targeting in 1992, proceeded by New Zealand which pioneered the strategy of inflation targeting in 1990 (see Haldane, 1995 for detailed insight and discussion on inflation targeting).

4.

The net benefit of depreciation can only be positive if the elasticities of export and import sum up to a value greater than unity, i.e. Marshall-Learner condition (Bahmani-Oskooee Ratha, 2004).

5.

Exports and imports of goods and services individually accounted for around 30 and 32 per cent of GDP, respectively, in UK (Office for National Statistics, 2016).

6.

Canada, France, Germany, Italy, Japan, the Netherlands, the UK and USA.

7.

They also used trade-weighted bilateral exchange rate volatility measures, in this study our exchange rate measure i.e. the Real Effective Exchange Rate is also a trade-weighted measure.

8.

Please see Primiceri (2005) for an interesting insight into the development of the TVSVAR framework.

9.

The modelling strategy entails modelling the coefficient process in (4) and one to one mapping from Equations (1) to (4) justifies this approach.

10.

The standard deviations σt are assumed to evolve as geometric random walks and classed as stochastic volatility, this is an alternative to the ARCH models with the crucial difference that the variances generated are unobservable components. On this aspect, Shepherd (1996) provided a good overview and comparative analysis of stochastic volatility models with ARCH.

11.

A point to note here is that none of the restrictions on the structure of V are essential as all the zero blocks could be substituted by non-zero blocks, with only small modifications of the estimation procedure. Nevertheless, there are at least two reasons suggesting a choice of V as the one described in the equation (8). The first one is related to the already high number of parameters of the model. Adding all the off diagonal elements of V would require the specification of a sensible prior, able to prevent cases of ill-determined parameters. The second reason is that allowing for a completely generic correlation structure among different sources of uncertainty would preclude any structural interpretation of the innovations.

12.

In addition to that, there are three reasons to choose and prefer Bayesian over the classical estimation approach to estimate subject class models. First, if the variance of the time-varying coefficients is small, the classical maximum likelihood estimator of this variance has a point mass at 0, related to the commonly called pile-up problem. Second, classical maximum likelihood is related to the high dimensionality and nonlinearity of the problem which is a problem as such a complicated model often has a likelihood with multiple peaks, some of which is uninteresting or implausible regions of the parameter space. Moreover, if these peaks are very narrow, the likelihood may reach particularly high values, not at all representative of the model’s fit on a wider and more interesting parameter region. In a Bayesian setting, the use of uninformative priors on reasonable regions of the parameter space is nevertheless effective in ruling out these misbehaviours. Lastly, the practicality of approach, writing up of the likelihood of the model is possible (at least in principle) yet, it is a hard task to maximise it over such a high dimensional space. Bayesian methods deal efficiently with the high dimension of the parameter space and the nonlinearities of the model, splitting the original estimation problem into smaller and simpler ones. Concomitantly, Bayesian seems the appropriate approach to consider.

13.

The Strategy of considering the whole sample and then discrete break has an innovative aspect and will give us further insight into the under analysis relationship in two different ways.

14.

For details on identification and structural interpretation, see Primiceri (2005).

15.

Just before the Northern-Rock, secured help from the British Government to over-come liquidity crises.

16.

The J-curve is the notion that describes a countries trade balance following a depreciation/devaluation of currency. The trade balance initially worsens before brining longer-term positive effects (for detailed discussion see Bahmani-Oskooee and Ratha (2004).

## References

Ali, S.Z. and Anwar, S. (2016), “Can exchange rate pass-through explain the price puzzle?”, Economics Letters, Vol. 145, August, pp. 56-59.

Aristotelous, K. (2001), “Exchange-rate volatility, exchange-rate regime, and trade volume: evidence from the UK-US export function (1889-1999)”, Economics Letters, Vol. 72 No. 1, pp. 87-94.

Bahmani-Oskooee, M. (1991), “Effects of exchange rate variability on inflation variability”, World Development, Vol. 19 No. 6, pp. 729-733.

Bahmani-Oskooee, M. (1991), “Is there a long-run relation between the trade balance and the real effective exchange rate of LDCs?”, Economics Letters, Vol. 36 No. 4, pp. 403-407.

Bahmani-Oskooee, M. and Harvey, H. (2015), “The J-curve: evidence from industry-level data between the U.S. and Indonesia”, The International Trade Journal, Taylor & Francis Journals, Vol. 29 No. 2, pp. 103-114.

Bahmani-Oskooee, M. and Hegerty, S.W. (2010), “The J- and S-curves: a survey of the recent literature”, Journal of Economic Studies, Vol. 37 No. 6, pp. 580-596.

Bahmani-Oskooee, M. and Kovyryalova, M. (2008), “The J-curve: evidence from industry trade data between the US and UK”, Economic Issues, Vol. 13 No. 1, pp. 25-44.

Bahmani-Oskooee, M. and Ratha, A. (2004), “The J-curve: a literature review”, Applied Economics, Vol. 36 No. 13, pp. 1377-1398.

Bahmani-Oskooee, M., Halicioglu, F. and Ghodsi, S.H. (2016), “Asymmetric effects of exchange rate changes on British bilateral trade balances”, 18th Annual Conference of ETSG, Helsinki, 8-10 September.

Bahmani-Oskooee, M., Harvey, H. and Hegerty, S.W. (2013), “Empirical tests of the Marshall-Lerner condition: a literature review”, Journal of Economic Studies, Vol. 40 No. 3, pp. 411-443.

Balcilar, M., Gupta, R. and Jooste, C. (2016), “The dynamic response of the rand real exchange rate to fundamental shocks”, Journal of Economic Studies, Vol. 43 No. 1, pp. 108-121.

Boyd, D., Caporale, G.M. and Smith, R. (2001), “Real exchange rate effects on the balance of trade: cointegration and the Marshall-Lerner condition”, International Journal of Finance & Economics, Vol. 6 No. 3, pp. 187-200.

Calvo, G.A. and Reinhart, C.M. (2000), “Fixing for your life”, NBER Working Paper No. 8006, Cambridge, MA.

Canarella, G. and Miller, S.M. (2016), “Inflation persistence and structural breaks the experience of inflation targeting countries and the USA”, Journal of Economic Studies, Vol. 43 No. 6, pp. 980-1005.

Chiu, Y.-B., Lee, C.-C. and Sun, C.-H. (2010), “The US trade imbalance and real exchange rate: an application of the heterogeneous panel cointegration method”, Economic Modelling, Vol. 27 No. 3, pp. 705-716.

Cogley, T. (2003), How Fast Can the New Economy Grow? A Bayesian Analysis of the Evolution of Trend Growth, University of California, mimeo, Davis, CA.

Cogley, T. and Sargent, T.J. (2003), Drifts and Volatilities: Monetary Policies and Outcomes in the Post WWII US, New York University, mimeo, New York, NY.

Edwards, S. (2007), “The relationship between exchange rates and inflation targeting revisited”, in Miskin, F.S., Schmidt-Hebbel, K. and Loayza, N. (Eds), Central Banking, Analysis, and Economic Policies Book Series, and in Schmidt-Hebbel, K. (Ed.), Monetary Policy under Inflation Targeting, 1st ed., Vol. 11, Chapter 10, Central Bank of Chile, pp. 373-413.

Forbes, K. (2015), “Much ado about something important: How do exchange rate movements affect inflation?”, 47th Money, Macro and Finance Research Group Annual Conference, Cardiff, 11 September.

Gagnon, J.E. and Ihrig, J. (2004), “Monetary policy and exchange rate pass-through”, International Journal of Finance and Economics, Vol. 9 No. 4, pp. 315-338.

Gerlach, S. (1994), “On the symmetry between inflation and exchange rate targets”, Economics Letters, Vol. 44 Nos 1-2, pp. 133-137.

GFCI (2016), “The global financial centres index 20”, available at: www.longfinance.net/images/gfci/20/GFCI20_26Sep2016.pdf (accessed 9 January 2016).

Ghosh, A.R., Qureshi, M.S. and Tsangarides, C.G. (2013), “Is the exchange rate regime really irrelevant for external adjustment?”, Economics Letters, Vol. 118 No. 1, pp. 104-109.

Haldane, A.G. (1995), “Targeting inflation”, A Conference of Central Banks on the Use of Inflation Targets Organised by the Bank of England, Bank of England, London, 9-10 March.

Hume, D. (1742), Essays, Moral, Political, and Literary, Kincaid, Edinburgh, A.

Keynes, J.M. (1938), J.M. Keynes to R.F. Harrod, 4 July, CWK XIV: pp. 296-297, available at: http://economia.unipv.it/harrod/edition/editionstuff/rfh.346.htm (accessed 31 March 2018).

Lane, P.R. and Milesi-Ferretti, G.M. (2002), “External wealth, the trade balance, and the real exchange rate”, European Economic Review, Vol. 46 No. 6, pp. 1049-1071.

Mishkin, F.S. (1998), “The dangers of exchange rate pegging in emerging-market countries”, International Finance, Vol. 1 No. 1, pp. 81-101.

Pattichis, C. (2012), “Exchange rate effects on trade in services”, Journal of Economic Studies, Vol. 39 No. 6, pp. 697-708.

Pourroy, M. (2012), “Does exchange rate control improve inflation targeting in emerging economies?”, Economics Letters, Vol. 116 No. 3, pp. 448-450.

Primiceri, G.E. (2005), “Time varying structural vector auto-regression and monetary policy”, Review of Economic Studies, Vol. 72 No. 3, pp. 8214-8852.

Rose, A.K. (1990), “Exchange rates some evidence from and the trade balance developing countries”, Economic Letters, Vol. 34 No. 3, pp. 172-275.

Rose, A.K. (1991), “The role of exchange rates in a popular model of international trade: does the ‘Marshall-Lerner’ condition hold?”, Journal of International Economics, Vol. 30 Nos 3-4, pp. 301-316.

Rose, A.K. and Yellen, J.L. (1989), “Is there a J-curve?”, Journal of Monetary Economics, Vol. 24 No. 1, pp. 53-68.

Shephard, N. (1996), “Statistical aspects of ARCH and stochastic volatility”, in Cox, D.R., Hinkley, D.V. and Barndorff-Nielsen, O.E. (Eds), Time Series Models in Econometrics, Finance and Other Fields, Chapman and Hall, London, pp. 1-67.

Sims, C.A. (2001), “Comment on Sargent and Cogley’s: ‘evolving post World War II U.S. inflation dynamics’”, in Bernanke, B.S. and Rogoff, K. (Eds), NBER Macroeconomics Annual, MIT Press, Cambridge, MA, pp. 373-379.

Sims, C.A. and Zha, T. (1998), “Bayesian methods for dynamics multivariate models”, International Economic Review, Vol. 39 No. 4, pp. 949-968.

Smith, M. and Kohn, R. (2002), “Parsimonious covariance matrix estimation for longitudinal data”, Journal of the American Statistical Association, Vol. 97 No. 460, pp. 1141-1153.

Wimanda, R.E. (2014), “threshold effects of exchange rate depreciation and money growth on inflation evidence from Indonesia”, Journal of Economic Studies, Vol. 41 No. 2, pp. 196-215.

Yildirim, Z. and Ivrendi, M. (2016), “Exchange rate fluctuations and macroeconomic performance evidence from four fast-growing emerging economies”, Journal of Economic Studies, Vol. 43 No. 5, pp. 678-698.