Capital Flows and Exchange Rates: A Quantitative Assessment of the Dilemma Hypothesis
Discussant
Tatjana Schulze (IMF)
Chair: Ozer Karagedikli (Asian School of Business)
All right uh thanks so much for excuse me thanks for having me here thanks for having me here um so uh it’s a pleasure to be presen in this paper uh which is jointed work with Andre Ferrero and Shan Shang Lee uh from Oxford University and University of
Liverpool uh respectively so um this paper is about a topic that you know we’ve been talking a lot about in the last couple of days and you will see that um I’m going to take an approach that is almost at the opposite side of the SP spectrum of what dimma just did
You know a very elegant and very insightful model I’m going to go with the butcher approach of having a very heavy uh and inelegant model uh but hopefully complementary to some of these uh uh discussions we’ve been talking about so the the motivation I guess you
Know it’s kind of trivial to this crowd but the way in which we’re thinking about this this paper is in relation to the transmission of monetary policy shocks and obviously what happened in the last couple of years uh with the tightening Cycles in advanced economies
And in particular in the US just uh got uh many people also outside of this crowd interested in this kind of question and the typical framework that we use to think about the transmission of monetary policy shocks is uh the uh often relies on the notion of a trilemma uh which was uh
Challenged uh a few years back in uh in uh uh Jackson Hall paper by Helen uh which basically consisted in her critique of this of this view consisted in two main ingredients the first one was an empirical observation that asset prices and capital flows tend Tove across countries and there is a very
Strong common component driving those uh those variables but then there were some very important the second aspect of her challenge uh was relating to the policy implications of the trilemma and in particular in her Jackson wpaper Helen made the case for uh or posed the question whether flexible exchange rates
Are actually uh enough to provide insulation to external shocks or shocks that drive this Global Financial cycle and also made the case for the potential usefulness of additional instruments to achieve macroeconomic stability okay um and so uh these are the two questions that I have in this bullet and what we
Want to do in this paper really is to try and revisit these questions that have been studied a lot in the last few years uh in in a way that as I said I hope is going to be complementary to many of the uh the studies that we’ve
Seen in the last couple of days and in previous work and we’re going to do this in an estimated open economy DSG model so it’s a model that is going to have the key features that are going to allow us to replicate the Global Financial cycle evidence and these key features we
Are going to which are going to describe in a bit more uh detail in the next slides basically rely on the notion of a dominant currency both in trades and in financial markets now more specifically what we do in the paper we proceed in three separate steps H in the first step we
Estimate a panel V where we estimate the response of macro and financial variables to a US monetary policy shock we’re going to focus on the typical small open economy okay so there’s been a lot of work as we’ve seen uh also in this in this conference on the transmission of us monetary policy
Shocks here we we’re going going to focus on small open economies with a flexible exchange rate which is going to key uh is a key aspect to answer the question that we have in mind and we’re going to show that uh in line with previous evidence as well that the
Demand or financial channel of monetary policy transmission dominates over the expenditure switching effect which is kind of a manifestation if you like uh of uh the notion of a Global Financial cycle being uh being at play after having estimated our panel V we’re going to um write down a two country DSG model
That we’re going to bring to the data by and we’re going to estimate by matching the impulse responses of the V and as I mentioned we’re going to add frictions in international financial markets as well as in trades and I’m going to show you how these frictions we’re going to
Add are going to be absolutely crucial to match the evidence that we have in our panel V in our empirical um part of the paper and finally once we have our estim DSG we can play around with it and think about policy experiments and in particular we’re going to uh do a few
Counterfactuals where we’re going to compare the benefits of having a flexible exchange rate we’re going to vary the degree of exchange flexibility in in the model and see what that implies for real outcomes the volatility of macro variables and Welfare and we’re also going to consider
Again in line with the lpaper uh Jackson Hole whether the addition of some instruments macro prenti instruments or Capital flow management instruments can mitigate the consequences of the Global Financial cycle we’re going to also consider a case in which uh of of an economy which has a packed exchange rate and is going
To introduce this tool these macroprudential tools and try to compare these various cases okay so that’s really what we want to do we want to exploit kind of you know the insights from the empirical evidence the Insight from a a relatively large DSG model that we estimate and then run these policy
Experiments uh in our estimated DSG of course we are uh building on a very very large literature uh on on this topic um most of you probably can spot your name in this list which is which is definitely complete and by for sake of time I’m going to skip the slides um and
Move directly into the the empirical evidence so as I mentioned we have a panel of uh countries we’re going to focus on 15 uh economies with a flexible exchange rate the list of countries you have in the first bullet so as you can see we have a mix of of advanced and
Emerging economies in in the sample and as I mentioned already here what we really focus and what is key for our analysis is to select countries that have a very flexible exchange rate so what we do we go to the uh um ilesi Rina rogo classification and we select the
First 15 countries uh in terms of the their classification of exchange of flexibility and we play around with the uh different set of countries in robustness analysis we’re going to consider a set of variables that is the most we think the most parsimonious way to characterize the effects of a Global
Financial cycle so we’re going to consider for the US uh a monetary policy surprise a variable that measures uh there is that is linked at least to Financial frictions in uh in financial markets or the on on the side of intermediaries the exas premium in particular and a measure of of real
Outcomes uh real GDP while in the domestic small open economy we’re going to consider real GDP CPI exports the policy rates the nominal exchange rate and another measure of there is link to financial frictions which is uh corporate bond spreads um at the in the local economy we’re going to work at
Month3 frequency and we’re going to have a sample that runs from 97 till just before the covid crisis really what what constrains here our sample period is the presence or the availability of credit spreads uh which is uh if you want or if you like you know the kind of the only
Addition that we that we bring to this literature given that there is a really a lot of work on the transmissional monetary policy shocks uh internationally so we construct this corporate bond spread starting from Micro Data uh and from at the country level and so and these data are
Available only from 97 uh but again uh we do robustness with uh different sample periods dropping um corporate bond spreads and as uh syv and Len uh do in in their restart paper we interpolate macro Series in this case real GDP and exports to the monthly frequency in our
Baseline but again we provide a lot of robustness exercise uh where we use other measures um in the paper okay so how does our pan V looks like um I mentioned we use a monetary policy surprise in our uh Baseline we uh use the monetary policy Surprises by um
Jinsky and karadi which again have been discussed already quite a lot in the past few days we’re going to plug these surprises directly into our specification of the V so we’re going to use them as an internal instrument in our V and uh as you can see there are
Two blocks of the VR here there is a US block that contains the monetary policy surprise the exess B premium and US GDP and the small open economy variables which are denoted by ni subscript and we have uh GDP CPI the uh exports the policy rates the exchange rates and
Credit spreads as I mentioned we we we we picked a relatively parimon specification that we can then we then expand the robustness now going back to some of the points we discussed yesterday this was for Lana’s paper uh what we you know given that we are focusing as I mentioned on a relatively
Diverse set of countries we have advanced and emerging economies there is a risk of uh uh basically the slope the dynamic slope coefficients I.E these coefficients um f here to be quite heterogeneous across countries which could lead to some biases if you were to pull all these
Data into a fix effects panel and so what we do we use the mean group estimator which instead deals with this uh uh with this problem and that basically means we run this oil with these V country by country and then we take a simple average to construct our
Panel estimator and we can construct confidence and conf confidence intervals uh in a way that is described in in this paper by pe and Smith so this is what we get out from our um from our panel V I’m going to just guide you through uh what we uh the
Way in which we interpret these responses starting from what happens in the US so the top row here is the plots the response of us variables and you can see that uh even though it’s kind of a very parsimonious way to represent what monetary policy shocks do within a given
Economy you can see that there is a very standard US transmission a tightening in policy leads to an increase in the excess bone premium uh in this case note that we normalize the shock such that it leads to a 25 basis points increase in the uh in the EVP this is roughly
Speaking a two standard deviation shock um in the work for example by ger and karadi just to give you a sense of you know the magnitud of the shock that we use so there is a tightening in financial conditions and there is a fall in economic activ ity with a classic H
Shape uh and magnitudes that are in line with previous estimates now let’s move to what we think is the most interesting bit of the analysis which is what happens to the local economy the small open economy let me start by highlighting that uh there is again some something standard that would you would
Expect happening which is a depreciation of the nominal exchange rates in the small open economy but then there are things that are you know at least if you start from the notion of of the framework in the trma there are less less or if you start from a mundian
Framework are less standard the first thing that we note is that there is Comm movement of corporate bond spreads across countries and this is a first manifestation of financial frictions being at play as well as a Global Financial cycle being at play given the co movement and the tightening Financial conditions across
Countries the second thing is that the expeditor switching effect doesn’t seem to dominate over a demand or financial channel of monetary policy transmission why do I say that well because despite the effects depreciation in this in the small open economy you have that exports do not increase indeed they fall
And quite uh severely if you look at the magnitudes and this leads to a fall in Real GDP and you can see that what happens to CPI inflation overall you know you have two opposing forces here you have the imported inflation going up because of the depreciation but you have the demand
Or financial Channel pushing down instead on CPI inflation which leads to a very flat near zero and very insignificant response think of this remember think of this as the average country in our sample okay so this would be the average response and that would be the average response of CPI uh which
Is in line with the uh with this strong response of real GDP and exports and finally how does the small open economy response in terms of the policy rate you have a very tiny uh tightening of monetary policy which is barely significant and eventually there is not
Much response which you can think in different way you can think that either the policy rate is just responding to the tiny blip in inflation that we have an impact or you can think of these economies having some form of response to the exchange rate remember these are flexible exchange rates but nothing
Prevents these economies to have some embedded response to exchange rates in a monetary policy rule so we think that this evidence is consistent with uh previous evidence on the Global Financial cycle hypothesis we stress the fact that these demands and financial amplification channels dominate over the expenditure switching
Channel to make sense of what we observe in in the data and as I mentioned already we do a lot of robustness exercises we consider larger sample of countries uh we also consider a longer sample period where we drop our variable of credit spreads and so on and so forth
And you know you can see in the paper that the results are uh or at least the impulse responses have shown are very very robust to all of these exercises right so given this empirical evidence I’m going to move now to the to Country DSG and uh we’re going to build on the
Work by Aoki Kaki and in particular the work by o ainan Albert carto we’re going to have um I’m not going to have time to kind of go through the model as I said you know this is on the opposite side of the spectrum of a simple elegant model we’re going to have
A lot of ingredients so what I’m going to try and do is to give a sense of the main ingredients that we have without going too much detail I hope you’re going to get you know the intuition for what are the important mechanisms at play so the household sector is going to
Be uh kind of uninteresting and symmetric across the two economies what makes the model kind of more interesting and more uh in a sense particular is an asymmetric International Financial structures we’re going to have Banks both in the foreign economy and in the domestic economy and the asymmetry comes
From the fact that foreign banks are going to raise fund domestically but they’re going to operate internationally so they’re going to lend both to their own foreign firms but also to the small open economy firms so think of these as being the hgim one or you know the us if
You like well home banks are going to be raising funds domestically but also they’re going to borrow internationally from this Global Banks okay so we can see that there is an asymmetric nature and we’re going to therefore think of the home economy as the small open economy we’re also going to have a
Relatively sophisticated uh produ prodution side of the economy we’re going to have four different type of firms uh and this in particular is going to allow us to as I mentioned before to match some of the Salient features of the Global Financial cycle in particular we’re going to assume an asymmetric
Pricing when it comes to International Trade we’re going to assume that in the home economy firms are going to adopt a local currency pricing Paradigm while this not is not going to be true in the US so the us is going to be producer currency pricing while there’s going to
Be local currency pricing at home and this is going to allow us to mimic uh the notion of uh a dominant currency pricing Paradigm as documented by uh previous papers and we’re also going to introduce a notion of impers imperfect pass through uh which I’m going to show
You is going to be important to uh to mention empirical evidence which means that the low one price from the point of view of the home economy is only going to hold at the dock so I’m an importer I’m going to import goods from you guys
You are you guys are the US and and I’m going to adjust prices only frequently so there’s going to be limited pass through from uh import prices into domestic CPI so in a sense as I mentioned before the the model is a model that has a dominant currency notion both in trade
And in finance so I’m going to try and give you a schematic idea of how the model works so we’re going to have a uh and this this going to be related to the financial flows in the model so as I mentioned there is a haimon which has
Global banks that fund themselves on foreign households and lend to domestic firms there is a similar structure in the receiving economy but what makes the model kind of interesting is this asymmetric nature of International Financial flows where foreign banks are going to lend to local banks so the key
Here is that these local banks are going to fund themselves from domestic deposits and international borrowing in interbank markets okay now an additional word on how these Banks work given that they play a a crucial role in the model these are modeled as a standard ger and karate um
Agent so the foreign banks are going to have their balance sheet fully denominated in dollars and they’re going to as I mentioned issue deposits to foreign households and then to both domestic firms and uh uh small open economy Banks the small open economy banks are going
To be uh more interesting so as we mentioned they’re going to fund deposits domestically and they’re going to also raise funds internationally B Star is going to be interbank uh borrowing from Global Banks and they’re going to use these two sources of funds plus net worth to L to
Small open economy firms now as in ainar ALU we’re going to assume that these Bankers can divert a fraction of their assets this is the gerter karadi part but this fraction is going to be increasing in the foreign currency share of borrowing over total Assets in other
Words the higher is my leverage in foreign currency the more I can run away with a fraction of my assets which basically means you know you can think of this as being that foreign funds are harder to recover than domestic funds and what is leads to is basically an
Endogenous uip wedge conditional on shocks that can drive a Global Financial cycle type of dynamic like monetary policy shocks okay so it’s really the combination of this balance sheet mismatch and the presence of a moral hazards on these intermediaries that is going to lead to this endogenous u
W again very quickly on how the production side of the economy looks like uh apologies this confusing but for those of you who know roughly how the model Works we’re going to have Capital producers which are going to introduce a role for the price of capital this is
Very standard these are our intermediate Goods producers uh which again are quite standard we’re going to have monopolistic retailers these were the the sticky prices and the LCP assumption is going to kick in and these are the importers which I mentioned already uh we’re going to assume some imperfect
Pass through based on the work by Tomaso monachelli in in an old paper so this is roughly how how the economy Works um I hope I give you a sense of how you know what are the relevant ingredients let me just now uh describe the policy Rule and
Then I can move to how we estimate the model so the monetary policy rule again is asymmetric uh and is a standard rule apart from one ingredient here so we allow the interest rate the policy rate set by the central bank to respond to exchange rate fluctuations
Potentially uh but again this is going to be a symmetric we’re going to be that we’re going to assume that this coefficient that governs this response is going to be zero in the US so in the US you can think of a very standard tailor rule while in the small open
Economy we’re going to allow this coefficient to be different from zero and in a sense we’re going to be allowing for a fully flexible regime if this F uh curly e is going to be equal to zero up to a up to a peg if you think about this coefficient going to Infinity
Okay and we’re going to estimate this model by doing impulse response matching and then play around with this coefficient in our counterfactual exercises we’re also going to introduce again in an asymmetric way some additional instruments and in particular this I’m going to describe them uh a bit
More carefully in the policy an is that I’m going to do in in in a couple of minutes again these instruments are going to be only in the small open economy only in the home country and we’re going to think of two different taxes a tax on foreign currency
Borrowing think of these as capital flow management measures or Capital controls and a tax on total credit uh in in the small open economy that you can think of as a mcro a standard microen uh policy um in the small open economy okay uh jump jum into the
Estimation so the way in which we proceed we are going to fix some of the parameters that we think are maybe less controversial in terms of um you know what we know from the literature and then we’re going to estimate the remaining uh parameters um which are
Kind of absent here and some additional ones which I think are interesting and important to match the evidence by doing impulse response matching okay so we’re going to try and match the impulse response we’re going to try and minimize the impulse responses of the DSG from the impulse responses of the
V um maybe the only thing that’s worth mentioning here is that we set the size of the home economy this is the small open economy 2.1 which is in line with the size of the counries that we use in empirical analysis and similarly we pick the home buyas in consumption parameter
Here uh to match the share of exports in GDP that is consistent with the countries in our sample over the sample Peri we consider so just to have a reasonably calibrated um you know part of the model when it comes to size and share of exports right so the impos matching
Procedure is quite standards we adopt a a basion approach uh so we have some priors on some of the parameters we minimize the distance between the black line which is the DSG response and the green line which is the V response taking into account the uncertainty around these uh um these mean estimates
Which are going to basically act as way in the impulse response matching procedure this is very standard so this is what comes out of the estimation as you can see we do a reasonably decent job and matching the Dynamics especially of the policy the financial variables cedit spreads the exchange rate and the
Interest rate we do a good job at matching inflation a decent job at matching real GDP and not a very good job at matching export even though we are still within you know reasonable confidence confidence intervals we’re still working on this aspect of the analysis and I actually had very
Insightful conversations with some of you yesterday on what may be going going on here uh but I’ll I’ll defer this to uh to chats in in the coffee so the over sense that we get is that you know we can you know match relatively well the the Global Financial cycle effects of
The monetary policy shocks that we estimate for the V now what I want to do um this is just you know how the estimated parameters come out maybe one thing to mention is that you know we get um if you compare our priors and the posterior you can see that the data is
Actually quite our impulse responses are actually quite informative about for the estimation which just let me skip this for uh uh for sake of time and let me quickly mention why the ingredients we introduced are important to be able to match the evidence um the first thing I want to do
Is to shut down Financial friction so I’m going to just kill the moral hazard ingredient in the model I can do it either in um the both the domestic uh the in only in the domestic economy or only in the foreign economy and the main
Thing I want you to take away is that Financial frictions are really crucial to a generate amplification that’s kind of uh you know maybe obvious uh they’re obviously important to match the evidence on credit spreads but also they’re very important to match the size of the exchange of depre depreciation
Because via the endogenous uip wedge that is caused by the financial frictions we get an additional amplification that feds into uh the Chang the second ingredient that I wanted to mention is the role of the asymmetric pricing uh assumption we made remember home firms do local currency
Pricing us firms do uh producer currency pricing the red line here shows what happens if we were to do producer currency pricing in the two countries and you can see that of course we wouldn’t be able to match the sign not even the size but the sign of exports
And this is kind of obvious because the LCP assumption makes exports not sensitive at all to fluctuations in exchange rate and you would actually go back to something that would be closer to a mandelia w where you have a big boom in export that eventually could
Even lead to an increase in uh uh in GDP in the home economy and a big spike in inflation when it comes to inflation here’s the third ingredient that I mentioned so if you remember we assume they’re imperfect pass through so we assume that a low of One Price holds
Only at the dock and this is the chart I wanted to focus on the red line here is what we would get if we were to remove the Assumption of imper imperfect past and as you can see you you without this assumption we would have that the strong
Depreciation that we get from the monetary policy shock is going to lead to an immediate spike in inflation which is clearly due to the depreciation and imported uh imported goods from the point of view of the small open economy so we think that this ingredient is actually very important to match the
Unresponsive unresponsiveness of inflation to uh to the shock now I have a couple of minutes left and so what I want to do in the last couple of minutes is to show you some experiments uh that we can do with our estimated DG the
First thing I want to do is to kind of go back to the original uh critique or challenge of a l to the TR hypothesis and think about the role of exchange of flexibility so the question here is whether or how much the exchange rate regime is irrelevant or relevant to
Shocks that generate Global Financial cycle type of Dynamics what I have here is a bunch of impulse responses that I get by varying the response of uh in the tailor rule to exchange rates just focus on the red dash line which is a currency pack so you let this economy to respond
Very strongly to movements in exchange rate which leads to no movement whatsoever in the nominal exchange rate and maybe not surprisingly to most of you going for a peg really in increases is the amount of volatility in the economy so in a sense you know the only
Point we want to make here if you look at the magnitudes is that the exchange rates regime is really not uh not irrelevant but what I think is the most interesting part of this uh of the policy analysis to think about additional instruments again I’m going
Back to Al lens kind of original Jackson Hall paper and thinking about okay if I introduce additional instruments can I achieve better outcomes in terms of allocation of or real can I achieve better real outcomes so I’m going to consider two taxes one tax on total credit which is basically reducing the
Total return on the portfolio of loans that banks in the home economy do and a second tax then instead is going to increase the cost of foreign currency borrowing from the point of view of the local bank okay and I’m going to specify a policy rule for both taxes that
Depends on total Credit in the economy okay so once I have these rules I can then put them at play and I can find I can we’re going to basically look and find a coefficient that maximizes households welfare uh conditional on the estimated uh version of the model and we
Can see how with this rule in place either the um macro the macro Pro rule or the capital control rule what kind of allocations we get so this is what I’m showing you here conditional on the macroprudential policy rule so remember this is a tax on total credit well in
This case is subsidy given that is is a negative shock and what you can see from the red line which is what we would get in Award with this additional instrument at play is that this macro prenti rule is actually doing a great job at stabilizing the economy so you can see
That uh the rule is going to which is acting directly on credit spread is going to compress credit spreads so it’s going to have a strong effect on the financial friction of uh of the economy it’s going to dramatically reduce the effect of the monetary policy shock in the US in the
Small open economy when it comes to GDP uh but it has a in a sense kind of a a very small side effect which is is going to lead to slightly higher increase in inflation which also leads to a slightly higher increase in the policy rate now can we how does this
Compare to the foreign borrowing the capital control rule well it turns out that these two rules are actually uh relative work in a relativ relatively similar way the the the capital control rule is also going to compress credit spreads it’s going to reduce volatility in outputs but differently from the mro
Potential rule there’s going to be a difference response in inflation here which mainly comes from the effect of on the exchange R and this kind of obvious if you think that the micro the um Capital controls rule is basically acting directly on the wedge on the uip wedge that causes amplification in
Nominal exchange rate but the bottom line I want you to get from these pictures is that you know you can think of these two rules these two rules as being kind of substitutes uh and Achieve similar similar outcomes last slide and then I’ll I’ll stop the last exercise we
Think of or we want to consider is a country that does a peg but also on top of it has one of these two rules at play I’m showing you here the macro prenti policy rule just for a reference the red dash line is what you would get with the
Currency Peg with no rule whatsoever so black is the Baseline red is a a word with a Peg and blue is a word with a peg plus the macroprudential rule at play with where again we find the coefficient in the rule that maximizes household welfare and this is quite interesting
Because you can see that you know the allocation that you get out of this experiment is not too dissimilar from the previous two so again you can limit a lot the response of output you can limit a lot the response of credit spreads but in a sense this comes at the cost of
Uh much higher volatility when it comes to inflation if you remember the previous numbers inflation was going up by less than .1% here inflation actually falls falls because the nominal interest rate increases and the inter the nominal exchange does not depreciate but these numbers are much larger and you know if
You compare the real GDP number as well you get um a much stronger effect yet it was surprising to us to find you know what I would describe as quite uh powerful effects from these rules in a world where you have a peg I’m out of time so I’m just going
To leave you with with basically this slide that summarizes what I just said um just a word on what we’re planning to do so as I said you know we want to understand a bit better the response of exports and we are also trying to do a
Bit more uh formal welfare analysis you know I’ve shown you a bunch of uh uh a bunch of input responses and you know we we’re looking at volatilities but you know then the next step in the paper is to be a bit more formal in terms of
Trying to assess the welfare effects of these different policy rules and with this I’ll stop and I look forward I’m looking forward to Tatiana’s comments [Applause] thanks all right uh thanks so much to the organizers for inviting me to discuss this paper uh today so in a nutshell the paper
Quantitatively revisits an old question so do flexible exchange rates inate from us monetary policy tightening and the paper Builds on three main building blocks in the literature um so following a US monetary policy tightening the classical mandle flaming um framework suggests that there is a trar so that flexible
Exchange exchange rates help with macro adjustment um uh while you can still um independently conduct monetary policy because there is an expenditure switching channel um such that an fx depreciation following the US monetary policy tightening actually boosts exports and and helps with the recovery then elen Ray building on her
Um Jackson Hall lecture says well there’s actually a dilemma so there’s no insul from the Global Financial cycle regardless of the exchange rate regime um and that’s because there’s a powerful Financial channel so that in the presence of strong financial spillovers from the core countries so from the US
To the periphery um there may be even a perverse effect where um a monetary contraction in the periphery country um is not um is actually expansionary so not contractionary um and so that’s because um Capital flows are both Pro cyclical and boring costs are counter cyclical and driven by the Global Financial cycle
And then finally the third building block is a more recent work by gadal on the dominant currency Paradigm um that suggests that there’s no to the little to no macro adjustment um from an fx depreciation because the expenditure sing channel is actually muted when Global um trade goes down altogether
Following a US dollar appreciation um ultimately how these channels operate is an empirical question and this is what this paper tries to answer in a quantitive framework and so it um uses on the one hand um a panel V with a classical y yarinsky karate shocks um cheski identification all very standard
Um and then um uh uses a 2count dsge model U that is matched to the V using some combination of financial frictions so crossb lending and nominal real frictions um incomplete exchange rate pass through and I’m going to refer to to local currency pricing as dominant currency pricing throughout this um
Presentation so some um DCP for the empirical fit um and just to mention the different to a standard small open economy model that has been used widely in the literature is that here in this model we have some um feedback from Financial fictions in the um core
Country as well so in the say us um and some foreign demand contraction that is not just model in reduced form and then um ultimately the paper tries to answer should we nonetheless stick with a flexible exchange rate regime uh by looking at some counterfactuals um what
Happens under a peg um and how can we introduce some taxes on domestic credit and foreign Capital flows um to improve on outcomes um so the paper Builds on a lot of literature I’m just mentioning a couple of papers here that um might be relevant as well that I was missing from the
Literature review um but I think the key contribution of this paper is really that it revisits the Dilemma in a quantitative framework that is really matched to empirical evidence cross country um and uses these three core uh frictions so uh including dominant currency pricing or local currency pricing and dominant currency financing
Um and considers Financial frictions also in the um core country in the hegan country and how they spill over through an endogenous uip premium um so I think it’s a very nice combination of a model and CrossCountry evidence and the paper is really well written and without much fuzz as I would
Say without like big belts and whistles um just briefly um the key results the model fits the data well uh the VR results are very robust and suggest that flexible exchange rate regimes do not offer enough insulation to the Global Financial cycle um so um the financial frictions incomplete pass
Through and local currency pricing are very relevant and uh explain pullovers and the financial Channel seems to dominate the expenditure switching channel so U mundle flaming in that sense doesn’t hold and then the coun the policy counterfactual show that fle flexible exchange rates uh and a domestic domestic macro proden policy
Measures I’m going to refer to mpms and cfms in the IMF terminology sorry for that but acronyms always help to squeeze it on one page um so they achieve the highest combination of welfare gains and limited mcro volatility um and in fact the um Peg increases macro volatility
Which is not surprising as we already heard from others because you forego um independent counter cyclical monetary policy but that can still be mitigated through a combination of um cfms and mpms of um taxes so just a couple of comments and Food For Thought um on the paper so
First of all um the paper considers cross country both EMS and AES um in the extension also um 24 economies but I’m wondering whether bunching those countries together actually veils a lot of heterogeneity the whole narrative is spun much more in the direction of emerging markets I would say um and how
They are vulnerable to the Global Financial cycle and we can see potentially the problem with that uh in the input responses in the VR so we have a nominal exchange rate depreciation but then we see very insignificant results on uh on inflation with very wide con confidence bands and in fact the
Literature tells us that um exchange rate pass through to CPI inflation may be heterogeneous across countries so Emerging Markets tend to have higher exchange rate pass through um higher currency risk Premia potentially also um exemptive fewer um CFM mpms in place whereas advanced economies have much stronger monetary policy Frameworks um
They are stronger inflation targeters have lower reserves and lower currency mismatches um so here some work by colleagues um at the IMF that have found that um markets actually have much higher pass through rate of um a nominal exchange rate depreciation um to Consumer prices and this is although there is a
Similar response of import prices so what umia called at the dock import prices um where where they find actually close to complete pass through of course there’s lots of evidence out there also some evidence suggesting that exchange rate pass through has decreased a lot over time but uh this is really like a
Sample of more than 60 countries over more than 20 years um showing some heterogeneous pass through so um looking also at the other two frictions in the model um DCP um so both advanced economies and Emerging Markets show um high levels of um currency invoicing in US dollars but
Still there is some heterogeneity in the sense that Emerging Markets have even more DCP and then also looking at um Financial fictions um we know from the literature that uip premium and advanced economies are an average zero over time and reflect more deviations from um full full information rable expectations
Whereas in Emerging Markets um they tend to be positive reflecting risk Premia so all of this is just to suggest that um bunching the countries together May Veil a lot of heterogeneity um and may therefore also have different implications for the quantitative results of the model um in
The sense that there there’s some mixed evidence um of where of the directions that inflation takes in the impulse responses and I think that’s important for the um policy implications because ultimately um the inflation channel will mean that an exchange rate depreciation will um have a strong real income effect
Um if exchange rate pass through is High um which will also imply less expenditure switching the policy rate responds very strongly to inflation more than to Output which will further um raise credit spreads and exacerbate Financial friction so I think it is important to distinguish a little bit
More between um Advanced and Emerging Markets um so one could estimate the VR um an exchange rate pass through separately for these countries and even also consider nonlinear exchange rate pass through so that you only have high pass through When U there’s a very large depreciation in
Essence then um a second um comment and I’m Bur already mentioned it so all of these um input responses from the model look very fine and match the data very well except for that one outlier here right exports um so is that maybe because of limited expenditure switching
Or because the financial channel in the model is actually not strong enough to lead to a big contraction I was wondering and I don’t know what your conversation were earlier um with others but I was wondering if maybe the the dominant currency pricing channel is not strong enough um and that
Could be tackled for example by um by expanding the model from a two country to a three country framework to really allow for a strong contraction in global trade such that um a US monetary policy tightening appreciates the US dollar and leads to a contraction and trade everywhere so that also the Third
Country which is essentially M mirror image of the Home Country we’ll see a contraction uh in in output and in trade domestic demand um which will then more likely um lead to such a protracted contraction and in exports and we can see that also from the gopad paper that
Um DCP in the three-count model actually leads to a large contraction um and ultimately I think that’s important because if it is in fact dominant currency pricing that explains in the data why flexible flexible exchange rate don’t bring about enough expenditure switching or trade um to counter the financial Channel then it is
Even a stronger case for um CFM mpms for um for taxes to Target the financial fictions in the model to ease monetary policy trade-offs um third comment I’m just going to keep it brief so the whole narrative is around the Global Financial cycle which is about large Cove
Movements and asset prices and we’ve seen that through the Cove Co movements and spreads and credit spreads um but the Global Financial cycle is also about credit Cycles so not just about risk off shocks but also risk on shocks um so to strengthen the uh the narrative a bit
More uh one could also consider what oops what happens in response to us monetary policy easing uh in the v would we see the corresponding flow of capital into small open economies as the Global Financial cycle suggest um and the VR could also be backed up by um by
Variables that speak more to the Global Financial cycle such as leverage uh and and credit contraction and capital flows rather than just credit spreads maybe that may also help you overcome the sample issue of a limited sample of credit spreads um finally on okay I’m going to skip this one because you actually
Answered that I was wondering if you are using um only pure monetary policy shocks or or a mix of of both so I’m going to skip that um and then finally on the difference between um credit taxes so macr credential policy measures in the model and capital flow income
Taxes um I noticed that the welfare gains under the mpms um are actually the largest um relative to micro volatility which is also what um what evidence more recent evidence suggests um but at the end of the day the capital the the tax on Capital inflows um also constrains banks
In uh Levering up in credit extension so could we find could we find an optimal level of the uh tax on foreign Capital inflows that would replicate the same welfare gains that the macr presential policy measure can achieve um so maybe just a little bit more discussion on how these
Two tools differ or how they are the same in your model or if they are essentially targeting the same externalities then maybe you could think about how you could design them in a different way so that they um Target um the externalities more explicitly all right I’ll stop here in
An interest of time so I think um this is a very nice paper really enjoyed reading it and yeah main suggestion is just to differentiate um more between Advanced and emerging market economies thanks so [Applause] much all right I’m going to take a couple of minutes to answer to these great points
Uh that Tatiana made um so on erogeneity so the first point was about erogeneity so as I said we’re mixing Advanced and emerging economies uh I’m sympathetic to the comment you know we can try and do it for advanced economies only uh I wouldn’t be able to do it for emerging
Economies only because I want flexible exchange rates and I wouldn’t be able I need a large panel to be able to estimate carefully the imun group so I think I’m sympathetic to the comment the only thing I want to mention is that you know the model is going to be able to
Match given these three ingredients my hunch is that the model is going to be able to match any kind of permutation of uh uh of the average country and so you’re right that our country at the moment is a bit of a weird Beast because
It mixes some stuff but you know if you if you were to consider a different average country the mod I will have a very good shot and matching whatever whether the implications will still hold it’s a good question and so that’s why I’m sympathetic to the comment so that’s
Something we will definitely do um on exports I really like the comment it goes actually similar lines to what I’ve been discussing with others um you mentioned you know it could extend so the the issue here is exports do not respond as much in the model as in the
Data um Tatiana suggested we can go to a fully-fledged three-count w where the do and currency Paradigm is uh better specified you know in our setup we just mimic it uh you would need the three country setup to actually carefully describe it uh the other option would be
To write a model of roundabout production where you have you know some form of GVC you have intermediate inputs that you export to produce in both the UK in the domestic economy and the and the us and that would I think allow us and this was the suggestion that I was
Discussing actually with yorgo yesterday would allow us to give an additional kick the other option and I would be I really would like to hear what you guys think about it maybe over coffee the other thing we were thinking about was to you know there’s some evidence that working capital constraints are stronger
And more severe for exporting firms and so we’re thinking about adding that kind of constraint where you need to finance some of your you know wage Bill and your costs and this is kind of relatively strong stronger for exporters that would lead to a stronger fall in in
Exports um then you had a comment about the the Global Financial cycle um again I’m sympathetic to that you know we focus here on the on the monetary policy shock mainly because of the impulse response matching we have PR previous work where we actually looked more carefully at leverage uh actually this
Builds on previous work with Andrea and um alexanderi were you know we’re considering shocks to leverage of Global Financial intermediaries so there is a there is all that notion that we don’t touch one in this paper but you’re right that we can probably bring it up a bit
More and we can do that certainly the in the new version of the draft and then you know the the the the last point was about you know the the different policy rules and again this is very much work in progress um we’re still trying to work out exactly you know what’s going
On given that we’re getting these welfare numbers uh numerically I just want to clarify one thing because I think I went a bit uh bit quickly in uh in my presentation so the two taxes that we consider a tax on foreign borrowing and a tax tax on the
Total lending portfolio of the bank are going to act on two different wedges in in the model so the tax on foreign borrowing has a direct effect on the endogenous UAP wedge that arises when there is a shock in the foreign economy okay while the tax the mcro pro tax
Instead is going to act on the total credit spreads in the economy so I think eventually and this goes back to your pre to your first point I think it really depends the optimality is going to depend on whether you are a country where these uip variations are are much
Are very important maybe much more so than domestic credit spreads or the Opposites if you have instead that domestic credit spreads vary much more than uip uh variation so I think this is the margin what that that can make these two taxes preferable depending on the
Setup that you consider um but as I said this is kind of we’re still working on it and I think this comment is going to be useful to try and get to the bottom of it so with that you know thank you very much really useful comments uh and
I’m happy to take questions if there are any or if are all in need of coffee I’m very happy to discuss our coffee thanks um I think the paper is is very ambitious to challenge the trima um to do that I think you might want to consider to look at uh price stability
Versus macroeconomic stability uh as a as a measurement you know to to answer the question um my second um comments is or question I just try to understand the parameter attached to exchang rating your monetary policy rule in the domestic economy why you you vary this this parameter will give you floating s
Rate and and pack exchange rate to me is just the magnitude that whole policy instrument in response to changes in exchange rate Um uh another comments will be I agree with the discussion that uh macral um policy tax on Capital influence gives you uh or pretty different well very improving is something a little bit old I think you might want to look into it yeah thank you um are we collecting oh no so thanks
I in reverse order I mean through the L of the model the tax on foreign borrowing obviously is going to reduce volatility because it’s going to is going to act on a friction that generates amplification in the model so so intuitively from the point of view of
The model is going to just you know dumpen uip deviations which generate amplification in the model and so you are really acting directly on the main friction one of the main frictions that the model has so it’s not surprising to me that it leads to an improvement in
Welfare the surprising bit is whether you know it does lead to a better or a higher improving welfare relative to a macroprudential tax our results show that it’s not obvious that it is better or worse you know at the moment they seem quite it’s more of an equivalence result rather
Than a better and as I said this part is still in progress but I I I’m not sure I understand your confusion about why a tax can lead to welfare improvements but we can discuss over coffee for the monetary policy rule you simply you know if the exchanger depreciates you raise
Interest rates uh then it’s going to limit the depreciation um again it’s um and you know the more you do the more you’re going to limit the depreciation and in the limit you’re going to Peg your currency but again if this is not clear we can chat over coffee Um yeah uh I think it’s kind of I think we can clarify this it’s relatively standard I would say uh just a follow up on Tatiana’s comments onogen um I had the same feeling that um there are two variables at least where probably you have lots of heterogenity
Underneath it would be nice to see it if you could show kind of a chart with all of the responses and these two variables are uh policy response the interest rates and CPI and if you think of the advanced economies Germany uh Canada probably uh the UK with such a
Contraction in GDP you would expect an easing from policy um and now the question is why not and I suppose that’s due to the composition um and in our paper in fact we find that that advanced economy do e when there’s a a tight in us because in our case both uh output
And inflation contract and that’s also interesting that you find instead inflation does not respond because what would be my prior is that you would have a contraction of inflation downward pressure on prices through the old Channel this a commodity price channel for advanc economies and then smaller economies perhaps have upward pressure
Perhaps on balance you find this uh this neutral response and would be also nice to look at the difference between core and the headline probably uh if that channel I have in mind is is at play at all and there’s also a new paper from uh Benin and C that explored this
Role for oil in the transmission and also I have a final sort of question or comment what’s happening with inflation because the two charts look different sorry this okay uh let me clarify again in reverse order Um I should have mentioned this actually so these these impulse responses from the V are as I described the v a monthly frequency with the set of variables the impulse matching you know the DSG is a quarterly frequency and so what we do we basically bring this impulse responses
To a quarterly frequency to the to do the impulse response matching so if you look at the bottom panel here you’re going to see there is not identical to these responses here because here what we’ve done to convert a quarterly frequency in the right with the right Notions averages for financial assets
And end of period for real quantities to be able to do the estimation so that’s if you see a discrepancy between these response and the one in the v is that for example for inflation in the v we have a up an upper beep blip on impact which now is negative because we’re
Moving a quarterly frequency if you consider the what happens within the quarter in the V and this was needed to do the estimation now the other you made a point about and so you made a point about erogen uh following on On Tatiana’s point which again I fully
Agree with um the only thing I would stress here is that and I know that different papers and we have many of them you know if we were to put together a meta study of what monetary policy shocks do to small open economies the results I would say are a little bit all
Over the place and my way to think about it is that we have quite a lot of uncertainty around grouping which is the erogeneity point we have uncertainty coming from the kind of shocks that you use but crucially you also have erogena in the sample periods that you consider
And all I want to you know the main point I want to make is that these responses are perfectly consistent with a linear combination of the various channels that joanni described depending on how strong they are given the sample period given the sample of counters that you cons
Consider possibly given the shock that you consider so taking your exactly the same logic what would happen here is that you know the oil Channel or the commodity price channel is maybe not as strong as you know you would find you would maybe be in a part of the sample
Period where inflation is less responsive to movements in output you know we know that these things can matter we know that the slope of the Philips curve may have changed over time so all I wanted to point to is that what we have here is absolutely consistent with all these different channels
Working in many different directions you know it just work out the linear combination of those mechanisms and you could get this one last point I wanted to make is on oil which was one of my you know exactly where my first um the first thing that I that came to
My mind so the model has a notion of a very strong response of oil prices to uh to the monetary policy shock yet the response of CPI is absolutely identical to the one that we have in our Baseline and so that kind of let me think well maybe in this sample periods this
Channel was not as prevalent as in in previous analysis maybe for this set of countries this channel is not as strong and a similar logic can apply to you know the policy response and so on um but in general I’m very sympathetic to the coms because I’m very aware there is
Work out there that shows possibly different responses possibly a different ranking of how important these channels are and I think you know as people who care about these things we should just you know keep on working on these and get to the bottom of it uh but I I
Totally recognize that you know some of these things are different to what you may have in mind question