Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=

# Jumpstarting an International Currency

Authors:   
Discussant: None  
Video: https://www.youtube.com/watch?v=4do5ZhrqGwI&t=0s


## Talk (00:00:00 – 00:54:40)

[00:00:00] thank you everyone for logging in um it's really a pleasure to be here especially at this inaugural event so I'm hoping I hope to make it worth your while to log in but looking ahead I've seen the program of the number of speakers and they're really fantastic so  
[00:00:13] even if even if the level today isn't great the slope is amazing something you guys should come back whomever is online for the next few weeks um okay so this paper is done with Salim something will be in the Q&A in case there's clarifying questions you go along and it's called jump-starting an international currency  
[00:00:28] the starting point here so let me start with the before we jumpstart to the paper let me tell you our starting point the starting point of Salim and I was really reading Barry Eichengreen x' work  
[00:00:40] for now a few decades where he's been telling him through a series of books and articles he's told the story of how is that the dollar went from circa 1910 being 0% used outside of the US borders  
[00:00:55] to as you all know nowadays accounting for depending on the measure something like 70% to 90% of global transactions in the world and the story that Barry tells or documents and what happened is that you know you start in 1912 where  
[00:01:10] the dollar is already the world's largest exporter but all trade finance that involved exports or imports in dollars happened in London and was denominated in sterling okay and then  
[00:01:24] why do we pick 1912 or even better very big sign control because 9 to 12 of course to fame the the date of the Fair Reserve Act and the Federal Reserve Act not only delivered what we're usually used to talking about terms or factors through the founding of the Federal  
[00:01:38] Reserve it led to stable inflation stable interest rates stable exchange rates but on top of it and very importantly the third Reserve Act deregulate it significantly the US financial market in particular it  
[00:01:51] allowed us banks to open branches abroad so US banks and so far as a lot of the financing happen in the London market US banks could now open branches abroad starting therefore being able to supply  
[00:02:03] trade credit and potential denominated in dollar to all of these US exporters and importers are active international market moreover not only deregulated but Benjamin strong the first president the Fed Reserve Bank of New York was  
[00:02:16] extremely active and if you look at his memoirs he spends as much time with this as he spends with the whole inflation and interest rate stuff he was extremely active in the so-called trade acceptances market so what does  
[00:02:29] the Train acceptances if I was importing if I was a US company importing rubber from Colombia but from Chile or from somewhere then you know I often would want to borrow against that input before I was able to produce my output and sell  
[00:02:43] it and the way I borrowed it was by pledging as collateral the input that was on its way to me and then these were called the trade acceptances because this was collateralized if you want buy this bad input well trade stems were  
[00:02:57] very important for international trade and just again London was the key market but Benjamin strong was extremely active in Korean the regulatory framework as well as intervened directs in the market to create a liquid market where a bank  
[00:03:11] who has lent to a u.s. importer exporter could now go and sell that claim to other banks and interbank market or even to other fashion meters so much so was his intervention that the Fed had explicitly ins Purdue the ferries are  
[00:03:26] back in New York said we will be a buyer of last resort of these if their price if the interest rate if you borrow in dollars for export-import businesses gets too high and this was a very aggressive buyer last resort I mean there's some studies show that around  
[00:03:40] nineteen twenty four twenty five the Fed held as much as fifty percent of all trade acceptances so this is not just this last resort was a frequent resort when you think about the binder that was doing it so that's what the Fed did by  
[00:03:55] normal in a very short period of time by 1925 the dollar is already very large again depending on the measure or somewhere between 10 and 30 percent of credit was already happening in dollars and of course by 1945 it is the dominant  
[00:04:09] currency that it continues to be to today question was this because of these policies the Federal Reserve Act I highlighted was it luck it just was luck insofar as you know a war broke in London and that had a huge effect until international market  
[00:04:23] or was it even just inevitable us was on its way to become the world the credit the main credit of the world and if you're the creditor maybe you impose the terms of which you're willing to lend it yet so this is kind of the background we were starting from the work of very  
[00:04:37] prominent international economic historians Suleiman I instead we've been we were looking at the China and we were looking at what's happened to the Remi me and here's a little bit of a very very brief story of the Remi be in the  
[00:04:49] last 11 years in 2009 or any of you that China was the largest its exporter was already the large world creditor and yet its currency was not used at all it wasn't used at all because given capital pros you really couldn't use it  
[00:05:03] but then if you look at starting in July 2009 the Chinese government specially the People's Bank of China start adopting a series of policies and let me just list some of those and they're purposely right next to the ones that the Federal Reserve did in the 1910s  
[00:05:17] free to see the parallelism it's turned by great deregulation you could now settle trade claims starting with this pilot scheme in July or nine settle them outside borers and now there are several places where you can settle claims  
[00:05:31] outside of China and relate to trade it came almost as importantly with creating a secondary market for the equivalent of trade acceptances and credit but generally for payments in our immediate this is of course the hong kong offshore  
[00:05:46] market that for rmb which really sprouts around 2010 11 in others it was of course the stable exchange rate through the dollar peg that has been there for a few decades and finally came also with the buyer of  
[00:06:00] last resort in this case to the PBOC swap lines why do I call it the buyer last resort and not to get into the details of other swap lines work too much but essentially because I'll talk about more than later but since you think about them what they do is if I'm  
[00:06:14] a firm in Thailand and I'm importing goods didn't inputs denominated in RMB well then my bank will I would want to get a loan from my bank in RMB well with  
[00:06:29] the PBOC swap line with the Bank of Thailand does is that my bank is comfortable that it can sell off flow if you want give to the to the tied to the tie central bank discredit and the test of the bank is willing to buy it  
[00:06:43] because it immediately then uses the swap line to send it to if you want get the RMB against it and from the PBOC okay so you see this is a direct intervention essentially to make banks  
[00:06:55] in Thailand in South Africa in whatever other country Argentina that assignees swap lines they can be much more comfortable in lending in R and B because they have now the spire of last resort via their central bank but  
[00:07:08] ultimately by the previous outcomes by 2016 the IMF included the RMB in the in its basket by 2019 it's a roughly the RMB is 2% the foreign exchange reserves so this these policies are better this  
[00:07:22] time was a successful time in surprising ambu and from zero to being arguably an international currency was it a coincidence that the same policies led again to a country jump-starting was it  
[00:07:34] just luck again was it yeah were these policies active in it that's basically what this papers about and this is what the next I guess forty minutes I'll be top telling you about okay well that's where we're trying to  
[00:07:48] what are we going to do I'm gonna make two contributions in this question we're gonna write a model where firms choose both the currency which they sell but important of their choosing also the currency of the credit they get in particular the trade they use for the  
[00:08:02] working capital the trade crane against their inputs we're gonna focus on writing a model where there's gonna be three thresholds and these thresholds are the thresholds such that a company in South Africa would like to indeed  
[00:08:16] sell in RMB as well as get trade crap get working capital credit in RMB and those thresholds are gonna be thresholds one on whether boring and RMB is both cheap and liquid here what I mean by liquid is  
[00:08:30] simply that the uncertainty on the interest rate that you face the bar on RMB is low enough second that the exchange rate between RMB and safe again I'm in South Africa selling to Nigeria  
[00:08:43] the exchange rate between the Nigerian naira and the RMB is stable enough and bird that the costs of the non-credit goods of my domestic input cost so gonna try my South Africa this would be in  
[00:08:56] Rand and the RMB is the covariance between those is high enough so there's gonna be three thresholds if you clear those three thresholds then companies in South Africa will want to start using the RMB further invoicing and the constraint down  
[00:09:11] crucially is the complementarity and some choices that if you choose more RMB for invoicing you want to choose more for working capital and these two are going to feed off each other and that if you want is the key economic meknes that we're going to be trying to introduce  
[00:09:24] you to the literature then importantly central bank policies that is essentially these policies over here that we wanted to make sense of they are going to what they do is they're gonna make they're gonna increase the liquidity they're gonna make the credit  
[00:09:39] boring in RMB cheaper the pulse of the PBOC in doing that that's gonna encourage more RMB borrowing that's gonna lead to more RMB invoicing and so on through the complementarity okay  
[00:09:53] that's what the mall is gonna be the contribution is these thresholds we're not gonna be going after the question of is the army gonna become a dominant currency or not we're going at the other extreme how do you go currency that goes from a model that we're why is the currency go from being zero used to  
[00:10:08] being positively used by firms outside of its border in their dealings not with your own country so again in my example South Africa said it tonight here okay so that's the first half of the paper so second part of the paper is gonna use  
[00:10:21] exactly these PBOC swap lines Versailles between or 918 why because it turns out that and the previous you didn't sign these 38 all the same date rather you know it's signed the one in South Africa  
[00:10:33] if I remember in 2015 but the world with Nigeria in 2018 so we're gonna essentially try to exploit the variation across countries in time of when did they get an RnB swap line both variation  
[00:10:46] come countries gotta swap these are MB some didn't but some gun 2018 715 711 and use every agency whether this policy indeed contributed through the lenses of the model to the rme being used by the  
[00:10:59] countries involved and the finding is already is gonna be that it did it was helpful these policies very consistent with again the writings of Barry Eichengreen for the US and consistent with our theory but using now the RMB  
[00:11:13] we're gonna find that indeed in a linear probability model the chances that South Africa use their and be went up by twenty percent on account of the swap line itself okay briefly on the literature a lot of really great papers on why the dollar  
[00:11:27] with different mechanisms becomes a dominant currency we're focusing on the other extreme like I said not why it comes to becomes a dominant but the compliment errors that make it go from zero to being positively used outside US borders we're gonna rely a lot of the currency invoicing literature which is  
[00:11:41] very much about given the marginal cost you want your price to essentially mimic in a way that make precise your marginal cost and a sender work right Charles angle and get a Gopinath of it's got your own where we go bond the key here here is that you're also trish entirely  
[00:11:56] the currency of your marginal cost is you choosing the currency of your working capital so you can have this two-sided complementarity that's going to be if you want the novel theoretical element here third we're gonna the lecture is very much emphasized again  
[00:12:09] the currency voicing we're gonna there we're now going to innovate at all the innovation is on this focus on working capital in here we're particularly close I think to two very sand papers from the banker josh'll settlements young-shin and Valentina Bruno and Aaron and  
[00:12:23] mallalieu again emphasizing the word and also relying a lot on so you worked in mightier majority or a or another's have done emphasizing more the finance side of currency denomination in our case  
[00:12:37] like in these papers the financing of the working capital but where is that literature is most emphasize the dollar we're gonna see hey to what extent are people actually borrowing and letting in RMB as well in this case in the JumpStart finally lots of papers and RMB  
[00:12:50] not know lots of papers but a few papers especially this great paper by this blue scientist McDowell on the policies of the PBOC we're gonna contribute actually a model and an empirical test to try and make sense of this some work on USD swap  
[00:13:05] lines me and Salim at work someone that let me just say I don't really want to talk too much about about soft ones unless people ask me the questions let me just say that they're totally different animals the dollar swap lines in the RMB swap lines there is there in some ways it's  
[00:13:18] like talking about mortgages versus collateralized former bonds they're both loans are both collateralized beyond but they're just very different in terms of what they do turns out um and then finally on the empirics a lot of work on currency choice using disaggregated firm level data for a country in a short  
[00:13:33] period of time we kind of grew the other way in that we're gonna have all the countries in the world not so our level observation is the country but we have the whole network because we're gonna use this data from Swift on papers across borders and we're gonna use  
[00:13:47] payments rather than invoicing so it's really gonna be the flows the payments they get sent on and off so for our question was important to have it because again there's no firm level variation that's relevant for us the relevant one is the cross country  
[00:13:59] variation and so that that's why the country makes sense and that's you okay I can stop you for questions just in case anything is come up no questions right now I think you can all head off for tea from that so that's kind of what  
[00:14:13] we're trying to do now model so let me present to you the model so keep an eye on this box this box is gonna keep on showing up so this is the model which is gonna be a continuing firms index PI J that's gonna be on the horizontal thing  
[00:14:28] here gonna be in the unit interval think about this again this is individual firms in South Africa each firm sells to a bunch of markets that's the vertical axis this is our African firm choosing that sells to Nigeria Angola Portugal  
[00:14:43] England Argentina as well so those are in the vertical axis importantly the South African firm sells also to the dominant currency D and and the rising  
[00:14:56] currency which you're gonna call R this is not a coincidence D $4 are for a beef instead of calling a rising versus dominant okay so all the other markets it sells are here in the vertical axis the actual axis that I equals 1 RI  
[00:15:10] equals zero this is gonna be the dominant existing dominant currency and this is the currency that the market matching the currency that isn't so used but may get used at some point okay so that's firms markets small opening up  
[00:15:23] now the firm each individual firm has to make a couple of crucial choices and so we're gonna into three periods the simple three-tier model to highlight the magnet the two crucial choices of currency are done in  
[00:15:36] period 0 and they are one do I want to borrow in the rising currency or in the dominant currency now we're borrowing here to have working capital right to get your inputs or production so we're gonna see that when you borrow in that  
[00:15:50] currency you're borrowing it to buy the inputs in that currency so you're gonna be choosing essentially the composition of your inputs which are going to be either X R or XD that's the novel choice in this paper standard choices you're  
[00:16:02] gonna choose LCP versus PCP but also the other option is rising currency pricing or dominant currency price okay you know the average interest rate you know the relative cost of input but what you don't know and what you'll only learn  
[00:16:16] appears one and that's the uncertainty is you neither know the exchange rate which is very relevant for this pricing but also you don't know whether by borrowing in at this period in dominant currency versus rising currency which  
[00:16:31] one's gonna be more expensive you're committing exactly the technology of financing for the next post you need to actually do the borrowing and you don't know exactly interest rate you got phase in this period you borrow to exactly again buy these inputs using the  
[00:16:45] technology of chosen because of the working capital you're then you're gonna sell your goods you're gonna pay your borrowing we're gonna repeat it at and issue it the product the second peer is completely mechanical so if you want you can collapse into period one by the way  
[00:16:59] we find it useful just so that and fix ideas but this is really the fact with two period model since all the uncertainty realized in period one note that the structure we're comfortable doing the two period because you could put this into an infinite horizon dsg  
[00:17:14] because essentially then this would mimic what we see in the working capital DSG and you change in literature first period shows the technology against the price then you do the borrowing which you pay after should be a morning and night kind of setup in those type of models that in principle  
[00:17:28] you could frame ours it but we're not gonna do that but want to do that okay so period 0 let me put some equations to it you're choosing the stick you're choosing a de according to this meaning you're you know you're gonna produce an  
[00:17:43] expert you're the che fur combining working capital X and other inputs out you borrow against X you don't need to borrow to pay for L but you are choosing exact or that X  
[00:17:55] this technology is relies on rising currency inputs or dominant turns inputs inputs that you're gonna have to borrow are against or inputs it you have to borrow D currency against in step and so if you choose eight equals one it means  
[00:18:10] that you're gonna have all your inputs are gonna be our inputs if you choose equals zero there are gonna be D inputs D currency inputs in between a mix of our versus D inputs if you want so that's what G technology choice is ADA  
[00:18:25] that I was telling with the amplitudes appeared zero extension what if you could choose to get our inputs but borrow using D currency against them we consider that in a robust a section it turns out there's a very strong force  
[00:18:39] that if I'm importing our inputs I want to borrow in our meaning if I need to pay for the inputs in our currency I want to do working capital in our currency as well so here it's assumed that the two are the same likewise you do an extension where we do just any homogeneous degree one  
[00:18:54] production function in case you guys get hung up next stand a choice do you want to set your price according to the if I again South Africa but I'm gonna choose it in the rand in the currency of my  
[00:19:07] expert market the Nigerian naira in dollars in RMBS knowing that if I set the price in my currency that means that for each unit I sell I get the P I chose but if I do it in the foreign currency then my revenue even though I set the  
[00:19:22] price in my local currency then in the local custom in the export good then I'm gonna have to multiply that by whatever the exchange rate to get the ones I have okay then likewise for domina exchange  
[00:19:36] rates here are gonna be exogenous this is where I said this is a small open economy if you want partial equilibrium model of these firm choices in particular all these exchange rate there's actually continued of lemon zest I are gonna follow a log normal  
[00:19:49] distribution so that we can focus on for a second moment we've also done this again on without the log normal assumption when for any general distribution and then we do everything with secular approximations and you get the same  
[00:20:02] results without these I'm gonna make one assumption for the presentation it's really not important this vlog round was pretty important or the second order that's how we're going to derive things this one is not I'm gonna assume that the rate of depreciation as well as the  
[00:20:17] variability of the dominant the rising currency are the same why because if they weren't then every Express will have an extra term that involves that obviously you want to borrow and price indication that wants to that's going to  
[00:20:29] appreciate more or the one that's less volatile we thought that was again this was such so obvious that it wasn't worth carrying that in every expression so again sets equal also for our empirics since the RMB is pegged to the dollar this is approximate when the  
[00:20:44] empirics so again we thought that generality wasn't really needed and so that's why we did okay then period one if you borrow in dollars that means you have to borrow Q and you pay one but if  
[00:20:58] you burn in the rising currency you're gonna have to pay epsilon and the absence the casting meaning borrowing in D is gonna have no uncertainty borrowing in R is that an insert interest rate or at least appeared what do you know what  
[00:21:11] that is but appears you didn't know it that obviously what matter is a relative we could have made the borrowing in D stochastic as well all that matters of relative so we're gonna focus instead on this G distribution I noticed the distribution of potential  
[00:21:24] cost of borrowing in the art currency um but at period zero and I chose that I was gonna go the AR way I didn't know exactly what interest rates there were okay I knew a new distribution the average the moments but not exactly  
[00:21:37] realization to pay for the inputs I need to pay for the working capital these are these rows so again I need to borrow to pay exactly this amount per unit of input and then I also need to pay for my non-credit capital I'm going to set  
[00:21:51] these working capital prices to be again their exhaustion as' and they are unknown we can make them stochastic it will make no difference at all important slightly more important zero to the propositions we're going to allow for  
[00:22:05] the non-credit input to be stochastic in particular like with all the other prices login normal so it's gonna have some covariance to the D currents in the air currents so again here I'm in South Africa I'm gonna have to pay for some stuff which if I imported them and chose  
[00:22:19] the art currency I'd paint our currency but I also pay for lots of other stuff the price of all that other stuff think of it as a measure of say producer price index or measure just input costs that combines both stuff I paid in my ran the  
[00:22:34] local currency but also I may have you know imports in dollars cuz I'm poor oil say and oils in dollars there's gonna be some variation of that with a dollar as well as with the are putting it all together what's our marginal cost of production  
[00:22:46] what - all of the stuff is paid in this W off of the stuff is the working capital if I choose a t equals zero then that's dollar what does that mean I need  
[00:22:58] to borrow one at the price one over 2g dollars to paint row per unit times SD so that I expressed our marginal cost in terms of myself in rent likewise if I do in the RB with a new  
[00:23:12] bit here that the price here has the smaller half okay then period - this is perfectly standard we're gonna have console assisity of demand so if I price at LCP I set this price this is the  
[00:23:25] demand if instead I price say in the rising currency then in units of the rising currency then I have to multiply by essentially the rising kurtzweil to the Nigerian currency which since R  
[00:23:38] these are the bylaw with the with my South African currency I need to do the ratio of them to translate that again we've done an extension for a general demand function that has strategic complementarity across sellers I'm not gonna cover that in these about 50 minutes  
[00:23:51] okay so then combining it what's the profits if I say to LCP and there's an equal to expression for all the other regimes I set my price in the Nigerian naira this is how much there's demanded that's my revenue as a result multiply  
[00:24:06] that by the rand naira exchange rate there's so much I collect in South Africa I pre the cost in my domestic currency times how much quantity I sold that cost is this expression that depends on my choice of the of working capital depends on the randomness and  
[00:24:21] the exchange rates on the costs and on my interest rate crucial this is the crucial expression in some ways in the following sense here again building on the shoulders of Charles angle get a copy natal get  
[00:24:35] Scotty Roberta regal bond I mean diamond look in many others in this literature the key problems that these firms are risk neutral like standard risk neutral profit maximizing firms but they would  
[00:24:49] like ideally perf information to have a constant markup over the marginal cost because of all this risk any deviation from the optimal you know theta over theta minus when markup is gonna mean lower profits therefore they want to be  
[00:25:02] choosing the currency denomination to try and match the currency of their sales with the currency of their costs so that they're as close as possible to that optimal desired market and it's  
[00:25:15] gonna be all about the riskier for the firm's to deviate from that and that's the problem that it's trying to solve this works in the solution okay finally what is the policies we want to choose  
[00:25:29] remember back to my introduction if you still remember we're gonna have all these deregulation policies what do we think about them well they're essentially shifting that the interest rate in the rising currency to a new till the distribution such that the old  
[00:25:43] one first Taurus the Casa dominates it that is they lower the variance they lower the average in barmen RMB the swap line is a just one example of these policies what does the swap line do it says that look you can always born our  
[00:25:57] and be by your bank by the central bank by the PBOC at a fixed cost Epsilon's swap let's call it therefore once you introduce what when what you do is for the right had the right tail risk of  
[00:26:10] borrowing in in in RMB gets truncated by the prince of the swap line okay so essentially it is a Purdue transformation particular way of again going to a distribution that involves  
[00:26:24] less risk in this resource gas of dominant sense okay so that's what the and the insight of the whole lender of last resort literature is even if this does not get used very often and indeed the swap line to the PBOC have not been you  
[00:26:38] very often it's not used to say determines where they're being affected their effectiveness comes like in any other land last resort by the fact they provide provide right till insurance but it's sometimes they do get used and so they can be very effective in exactly  
[00:26:51] triggering credit as Benjamin strong to the 1910s as the Bank of England discovered in the 18th century with Badgett and the whole lender of last resort literature so that's the model any questions on the model if not I'll proceed to the predictions yeah so  
[00:27:06] Pierre Olivier grinch I had a question me once you do could you say a bit more about the stickiness and borrowing decisions how hard is it to switch borrowing currency in reality or to put it differently if there's stickiness and borrowing decisions is there still a uip  
[00:27:21] condition in the background between the different interest rates and exchange rates right so this so in terms of our model we're not going to impose uip sessions in terms of the model it could be there or it could not be there crucial so that's just in terms of in  
[00:27:36] terms of clarification on the model in terms of the the important question here is should measure that we had not uip but say CIP we had CIP holding then and it held exactly then it wouldn't matter which currents you are borrowing because  
[00:27:50] you can always change it to some other currency whatever parents you want using using for contracts so that so CIP existed then this would be mute in terms of what currency you want to borrow if  
[00:28:03] there are CIP deviations then you do indeed makes a difference which when you picked but still if there's a positive CIP deviation even though I'm borrowing an RMB if I could use futures to turn that into sterling then that simply a deviation would be just the extra  
[00:28:18] borrowing cost in some ways now in the case of the RMB there's essentially there's very few it's extremely hard to borrow in forward markets against RMB or it's extremely expensive so you can think about this as the fact that if I  
[00:28:32] were to borrow other currents if I want to switch from RMB borrowing to GBP barring or something else this should be prohibitively expensive right and so you can think of our assumption that you're borrowing one or the others being those cost those epson's are either the direct  
[00:28:46] cost of borrowing in RMB or they're the cost of borrowing RMB bla or started the cost of borrowing in say pound sterling Plus then the extra CIP cost of borrowing turn per in that guarantee now that's a very very large  
[00:28:59] number so the fact though for at least the empirics it's as if they're essentially not a CIP condition or it has a very very high cost and so you would focus then on the on the costs on that end okay and it can we can come  
[00:29:14] back to this in the Q&A that's very important right it's the extent to which you can borrow in which currency so here think about because another way to think about the swap lines averse what they do is they put a ceiling on CIP you can think about that what they're doing is saying well even if you're  
[00:29:27] borrowing GBP but you want to convert that into an RMB how expensive is that now the people you'll see the facto is in this case is not putting a ceiling but is actually creating almost such a market because you can then use a swap  
[00:29:41] line to do that currency okay good model predictions manage my time okay model predictions first prediction is one straight from again it's cocky if you want  
[00:29:54] Gopinath rego bar okay but applied here and especially applied for this threshold problem and that is first imagine a firm that chooses RMB will working capital will that firm ever  
[00:30:06] choose DCP and the answer is no I mean I have my working capital in RMB I have my AIA paid for my other inputs and whatever currency that is the tableau I'm selling to Nigeria why would I ever  
[00:30:20] price anything in dollars okay I'd always prefer to price in RMB because if nothing else and matching my invoicing currency with my working capital cards second when do I do that as opposed to  
[00:30:32] making say the the again the Nigerian naira well I do that if the variance of the local exchange rate the naira exchange rate with the rent is such that sorts of the variance of that is  
[00:30:46] sufficiently high relative to the RMB back why if you're above this threshold capital Phi then I rather match the the the RMB cost American Capital su that my  
[00:31:00] revenue rather than master the demand that's coming from Nigeria so so that's why that that works out that way use of time let me skip the UM look if alpha Z this is immediate if alpha is  
[00:31:14] equal to one because in that case mean this always hold the factory goes with one because then your working capital is the entirety of your inputs so if that's inside of your input you want to go and match your revenues your sales to the cost of your inputs and so you always  
[00:31:28] choose RCP instead of LCD you can see that if alpha is equal to 1 this always hold but if not it's not a third when you prefer then RCP the PCP well you do it if well I myself confirm I have two  
[00:31:43] forces here I want a price do i price in the cost of my working capital or the cost of my other inputs well if my working capital and other inputs are correlated enough then I'm happy to price it in terms of the RMD and that is  
[00:31:57] if the Sigma's above this threshold Omega which obviously depends on how important the working capital is in the first place and that's on the valuable proposition what about for credit the new choice here is that the PC PLC P  
[00:32:11] type of analysis but what about the choice of Ada first result ADA's bang-bang you either all want all your working capital to be in our or in the D curves you never want in between would I know what you want in between well you don't want in between because  
[00:32:25] again you're a risk central firm you want to maximize your profits you want to go for whatever is the cheapest thing there is and so you end up with this very instead she comes from the quasi convexity of profit functions with respect to cost of the inputs when you  
[00:32:40] then choose one as opposed to 0 you will choose it if this the cost of boring an RMB according to this particular expression is low enough relative to the average interest rate cost the average  
[00:32:55] cost of the working capital inputs and this threshold side their threshold a very neat thing about the mold where you set it up is that it's the same for all firms it doesn't depend on the firm J  
[00:33:06] okay it depends on first if your let me go through the larger if you are selling a lot in RMB you want to also are more likely to on credit RMB  
[00:33:19] that's the complementarity second if the RMB market if China was a big market for you in the first place that is immediately a strong fundamental attraction pole for you to want to do this your cred and R&B in other words if  
[00:33:33] I'm already selling a lot to China and I'm pricing to China in RMB that puts regardless of whether to Nigeria pressing R&B or not that is in itself a strong reason for to do finally thirdly if the covariance between again my the  
[00:33:46] RMB in exchange rate and the costs of my other inputs is very high again that creates this force from you want to denominator in capital in RMB as opposed  
[00:33:59] if you then combine these two propositions the latter the proposition you end up with our key propositions for policies what happens if you shift the distribution of credit costs if through bread liberalize your financial markets  
[00:34:13] if through policies such as the swap lines or such as Benjamin or essentially just more generally buyer last resort if you lower the make this market more liquid in the sense of for sarcastic  
[00:34:28] dominance although you see already the key condition is whether this animal here on the left this expression Falls if that is so then let's work sequentially through what happens if I  
[00:34:39] was already choosing where capital in the art currency this has no effect on the choice of RCP this is the fact that here this conditions didn't depend on a  
[00:34:52] second for so first stop is if I was already selling if if I was already um getting capital in are then do does that change the way when a price in Nigeria  
[00:35:05] no Neil Sedaka firm but the direction keeping fixed the fact that I was already invoicing in some currency whatever that was in Nigeria for my South African exports if I myself  
[00:35:17] confirm this shift means that if I cross that threshold the firm's that do cross this threshold here for which the shift means this now below the beside those shift to RMB credit and then and only  
[00:35:30] then for those who shift of the RMB credit then as long as given that I shift its weight equals one the lemon applies if I'm above the Phi and Omega thresholds for those I will now switch  
[00:35:43] to our CP and now given that I've switched to our CP then we go back to B that's what should our CP lowers the side more makes me want to get more our credit which in turn leads to more our invoicing which leads to our credit and  
[00:35:58] that's where the complementarities kind of worked out okay so let me do this with my picture we started here South Africa is Salaf confirms a lot of them are invoicing partly in dollar in some markets some of them in the local currency they were invoicing in dollar  
[00:36:13] for Nigeria they're investing in pesos say to Argentine we lower we have the central bank policies the central bank policies leaved across the threshold that threshold is a badge a star some firms  
[00:36:27] now find it more find it cheaper if you want cross the threshold such that they want to now borrow in our currency we move now this vertically to the right of those then some of them precisely  
[00:36:41] because they're borrowing in are they want to now price in art and that that is this threshold that comes from the thresholds in lemma one so what you end up is the world in which and majority from South Africa are still having credit in dollar and are then pricing in  
[00:36:56] either dollar or pesos but some from South Africa now borrow for the word capital in RMB and some of those will now price their goods in RMB that's the vertical side while others price in  
[00:37:08] whatever look for service the area here is the use of their and be both in payments sent and in payments received okay now if however it was the case as the covariance with a non credit input  
[00:37:22] think about again the ppi the domestic cost was low same idea but instead of throwing RCP here you would have done PCP still note that the policy leads to an hour usage but now it's just to the  
[00:37:34] thickness here you're borrowing in are but you never price in are with the exception of the Chinese market if I were the Quran says I enough this purple thick line becomes instead the square  
[00:37:47] okay if in turn by the way that covariance rises well that's gonna push this to the right that J star to the right that's what the compañeros changes the IRS are and the area increases by  
[00:38:01] this amount right so that's what the promise mentality so model makes for empirical predictions what there are even though there's 200 currencies so we'll try to an occurrence in the world very few of them  
[00:38:15] international why this for most currencies you're above the thresholds you have a relatively volatile currency with other with exchange rates with other currents in the world your small here I am I'm in Nigeria I have mine IRA  
[00:38:27] I'm small in the input basket because I I'm not so in terms of the sales of British firms or of transfers or others and it's really expensive for you to get borrowing in Ireland because there isn't the depth of financial markets that  
[00:38:42] exist so if the Nigerian central bank introduces a swap line as you're in IRS they're so far to threshold that wouldn't matter at all no matter what policy you do you're just not gonna be here at castle but if you're the PBOC  
[00:38:55] and you already have a currency that's relatively stable with the dollar and the specular dollar is also stable with other currencies if you're ready a big exporter such that a lot of countries a lot of firms lots of countries already have are already paying for inputs that  
[00:39:10] are coming from China in RMB potentially and you have started the chinese hong kong market and on top of it you now lower the cross of credit with a swap line then you can get a jump start and you can get a bunch of these Nigerian  
[00:39:24] firms they want to use RMB credit and to invoice to the Nigeria's of the world in that's the first prediction second prediction if I sort countries - go various within the RMB and Nigerian  
[00:39:36] Angolan British and other input costs then those for which this is higher are the ones for which we'd see a higher policy on our usage they're the ones to which you will see more of our firms  
[00:39:49] when twin voice and to get priority and finally imagine diamond in in Nigeria and South Africa signs the swap line well because I saw swap line and I actually import benefit from that from  
[00:40:03] South Africa in Nigeria this Sigma RW is gonna rise that is if my neighbor signs of swap line because they start using more RMB me Nigeria I want to start doing more RMB invoices as well as well as RMB credit and so these are the four  
[00:40:18] predictions that including the model I'm gonna turn now to the data form let me present the data I won't stop it for questions just so that is we're about 10 to 15 minutes from me stopping and so I'll just do the empirical the data in the empirics and then we'll do the  
[00:40:32] discussion that's okay that okay Rob or were there a lot of questions now that's perfect just keep in mind data very simply here are the data on the swap lines both number and amounts growing 2009 no swap line has been  
[00:40:46] reverted so this is monotonically increasing how much these are pretty big by now very very big in terms of mass as big as say the maximum usage of the dollar swap lines are the Fed including in the last month um how are they signed  
[00:41:00] just so you get a sense there's time variation we ordered here so said that color is the ones that also come with more flows between China and those countries thousand and I there were only a few crucial one with Hong Kong to start the hunk of our market and this is  
[00:41:15] how it look this is how it has expanded over time right so now a lot of countries ECB a gigantic one Hong Kong but also speckies than Suriname and a bunch of others and I I I think I checked Nigerian South Africa should be here that's why I picked them for the  
[00:41:29] example so that's our first data we collected the date the exact date per month of when each one of these were signed across countries the PBOC makes us a little hard so it takes a little bit of work to collect this but again  
[00:41:43] I'm happy to share those dates or acted in the appendix of the paper then we're gonna use the Swift database so Swift essentially you you all know when you send money across borders almost always goes through Swift so if there's a messaging service between banks to say  
[00:41:57] reliable source a bank a wants to send to reliable Bank B we have monthly bilateral data for the entire world so essentially what I have is how many payments went in August 2015  
[00:42:10] between Nigeria and South Africa broken down by currency okay for every month over here this is for those interested you can get that data from Swift it's available to other researchers as well crucially for us I haven't broken down  
[00:42:25] by currency and I have it at a monthly frequency and for the whole network of the world so I can exploit this difference of when different countries adjusted or not okay so my key measure is gonna be the RMB  
[00:42:37] share in cross-border non FX payments sent and received per month per country just real briefly not effects is very important this data often is dominated by the closing of a series overnight repos in currency so  
[00:42:51] important that you look at the non FX payments and it's broken down by that otherwise they'll swarm here's the Agra data against also swap lines the RnB you can see jump started meaning used to be essentially zero and it goes to one  
[00:43:05] percent two percent roughly three percent by the end of our sample that it's used this is on average across the different countries now if you look at the first thing we did of course is plot the data here's to share their army payments and here I did an adverse class intend to 18  
[00:43:20] against it's a share of China trade obviously if you do a lot of trade with China Mongolia here you're gonna do a lot of RMB payments but what really struck us in this picture is some countries it's the 45-degree line some  
[00:43:33] countries you trade a lot more than they use RMB this is the fact they are B again is not a dominant currency for the dollar you have the opposite most countries use more dollars in the attributed dollar but crucially look how many currents are here in zero and so this is the share cross the period  
[00:43:48] essentially a lot of the action the data is that a lot of guys start from zero and change to positive okay so if you were to run a regression on shares which we did but I won't I think I won't show you I don't have time to show you the focus really is not intensive margin the  
[00:44:03] focus is they made the extensive margin all these countries they went from using zero to going to the right that's what dominates the data as a result we're gonna focus precisely on non rich countries because non returns  
[00:44:17] may have again financial that can be financial trade so that we can focus on this kind of working capital firms invoicing we don't focus on non rich countries and they're all the action is do you go from here to there that is do you start using the RFP at all okay so  
[00:44:32] here's what here's kind of if you want picture of prediction of the model if you want this is the median RMB share and cross-border payments this is month before and after country sign a PBOC  
[00:44:45] swap line and for the ones who signed it for those 38 and here is kind of the you know graphical way to see the paper the empiric sorry you didn't use the RMB you sign a swap line you start using the rme simple as that  
[00:44:58] the effect this seems persistent if anything gradual and delayed it seems that there's a clear extensive margins switch around here so the sense if you're kind of more of a graphical guy that's empirics it does seem that consist of a proposition to  
[00:45:13] you sign the swap line you start using the RMB okay so here you can stop you if you don't care about regressions or other but let's do regressions and things in particular there's one key conservative course and the key concern  
[00:45:26] is reverse causality meaning maybe something happened that mean that you want to use more RMB because that same thing that meant you want to use more RMB also meant that you want to sign a  
[00:45:40] swap line with China I don't know maybe your people in your country started really liking Chinese goods if they really liked Chinese goods they have to use RMB to pay for them at the same time because your country is now so friendly towards Chinese products you also called  
[00:45:54] up a central bank and say let's sign up a swap line even if you swap lines are totally useless for RMB usage that's really that the the key concern it's some unobserved factor drives both RMB usage and Alex can sign swap line we're gonna deal with this in a bunch of  
[00:46:08] different ways I mean that's all the econometrics for then adds to that picture first if the fact was common to the country forget just I'm more I like my country likes Chinese things both currency credit and goods more than  
[00:46:21] others then we'd have oh sorry it's not the other way around then we'd have some country fixed effect on the other hand if this is just the whole world switched to kind of lighting more Chinese things both swap lines and payments and goods then there would be dealt with the time fixed effect so we're gonna have the  
[00:46:36] time intention tricks of that more than that though imagine the fact of some region specific trend you to trade with China or some again political friendless China productivity in village of China we're gonna use the share of RMB in my  
[00:46:51] neighbors so I signed a swap line how much RMB was being used by them to see if that's what led to the taper next me let me use a bunch of measures listed here of how much I trade with China to try and control for maybe you signed a trade agreement with China and  
[00:47:06] that's the third emitter fact next let me do Chinese policy very important here is there an RnB clearing Bank are you measure member the aged investment bank or are you through the belt and road initiative all sudden getting a lot of money from chime invest this we thought could be a  
[00:47:20] very big deal maybe you just there's a lot of investing Roden Belt melting roads in your country that's why you signed a small plan and that's why you're having all these things so we're going to controls for that as well so we're gonna do essentially fixed effects and all these controls to try and deal  
[00:47:35] with that in our first approach and so here's regression it's a linear probability model with the fixed effects and with all the controls I listed and then this is a dummy variable it essentially switches on when you sign the swap line so this point 28 is if you  
[00:47:48] wanted the direct map to this picture here once you do all the controls you get a very solid if you want and so far as doesn't seem to change love with all controls 13% impact of signing a swap  
[00:48:01] line to you going from not using RMB at all in payments and to receive to the problem that we observe that you see there our second approach let me manage my time I'm gonna I'm gonna take five  
[00:48:15] more minutes okay organizer start it should finish it within a minute but I'm gonna take five minutes that's okay um second approach for the reverse causality by again maybe I should have spent more time on this but again we're trying to control for anything that we could think of there leads to an  
[00:48:29] admitted fact that leads to more payments that come from our M bees and that be leading to you just signing also swap line just as a political posturing measure to second approach so slope length almost always get signed when  
[00:48:44] there's a state visit with changing pingos in visits your country okay that's just how these things work you sign that together a bunch of under so our idea well GG being visited South Africa in April 2015 but the state visit the Nigeria's 2018  
[00:48:59] instead let's look at the variation let's say well even if there was a submitted factor the fact that you visit South Africa a year or three years before you visit Nigeria that often the state visits were not organized to sign the swampland often organized with a lot  
[00:49:14] of randomness I'm to do with agendas and calendars of state leaders let's use the time series variation of when the state visit happened to as an instrument to when you signed a swampland so the first stage is enormous which again just  
[00:49:28] reflects the fact that when you signed the swamp way when there's a state visit that's when they actually signed agreement but arguably your point is that the data which that happen that is somewhat exogenous to whatever other factors can affect and so if you do that  
[00:49:42] if you do now an IV regression using the timing of state visits you get if anything actually a larger effect of the swap line on sorry oh yeah adopt the swap line on the probability that you  
[00:49:56] have to pay third approach which is both an approach to the reverse causality as well as a prediction in the paper remember the papers predicted that if 9g if select assigns a swap line that in crete that raises the threshold that  
[00:50:09] means the nigeria now wants to use more rmb as well so let's do that arguably South Africa sign the swap line is with that's XR Janice to whatever is going on Nigeria Nigeria didn't control the fax after it shows the signs wrong so gonna  
[00:50:23] see whether not the fact that you send a soft line but the fact that your neighbor sign a swap line really use that as the variable in the linear probability regression to see what happened to your RMD payments including  
[00:50:36] again both for all countries also excluding countries and have the swap line so I'm now gonna be looking so I really like these ones I'm gonna look not an idea but I'll end Ola or should have check I think I'm gonna say  
[00:50:49] Mozambique Mozambique does not have a swap line with China this is the fact that the neighbors of Mozambique Angola Nigeria signed a swap line do we see Mozambique older son using RMB even though it wasn't big itself as then nothing with China at all as in sign is  
[00:51:04] wrong in the answer is yes you do you get as you expect the smaller fact is more predict but you still see that Mozambique starts using the RMB just because South Africa sign a swap one okay find a particular model if we  
[00:51:17] sort on covariances sore on the correlations in the RMB and the contribute PPI you get some of the effect that you would expect that is that the stronger is the correlation the for the high correlation versus the low  
[00:51:29] correlation you get a larger effect the sample Falls quite a bit though because our measures of this w which again we use PPI we only happen for a few countries we've done this excluding China we've done split in payments and received we've  
[00:51:44] done it using again shares in probability we've done it where you look at immediate effect the first 12 months this is after all of these are to see whether this is just being driven by China or where this is all kind of the smell of it was just political posturing the straw finds it nothing and the  
[00:51:59] results are consistent with the theory so to conclude in my remaining two minutes what is that we did here we try to study to what extent whether a currency not  
[00:52:13] becomes dominant but goes from being you know South African Rand nobody uses it unless you're having relations South Africa Nigeria Mozambique do not use the South African Rand when they don't shoot each other um how do you go from becoming a national currency and we  
[00:52:27] observed were in the last ten years the RMB moreover we observed it and because both we have greater data and because of the stop lines of Janani we can study it because so many of the polls that it did we're the same policy that the dollar  
[00:52:40] did we can start to learn about generally whether these hypothesis that economic historians and others are made about the tolerance in the 1910 2000s can we test those can we learn about the RMB can we learn about international  
[00:52:54] currencies in China we were the model emphasizing precise because we want to study these policies emphasizing financial markets and working capital credit why cuz these faults are directly about being able to borrow in the current or model the key force was this  
[00:53:07] complementarity Jim crema invoicing and these threshold results then you go from zero to positive once you cross thresholds so you get this in the same way that the dominant curves which are often as threshold so say you become dominant we have thresholds or when you  
[00:53:21] go from zero to positive if you want and they're all about this component talk to the credit in the invoicing such as the policy jump start what do the swap lines do there were just our application they're very nice cuz they give us this  
[00:53:33] exogenous this variation that allows us to test these things that are incredibly hard to test all these dominant currency or just furniture choice are very hard to test we get that exhausted that birds variation that allows us to test them I try to explain to you what these things  
[00:53:46] do is essentially remove right tail risk of R&D financing and that indeed increased by 3% of our world finally what does our paper empirics tell us about about will the RMB become dominant or not the answer is they tell you nothing  
[00:54:01] about that I mean you know our model of course you could now use it to see to what extent this if you want this gets so big that the red line goes here and the RMB overtakes the tea we drew this year you really since it's three percent  
[00:54:15] or 30 percent you should really have the red line more like here we were studying the emergence of this little rectangle over here and it should be much much smaller if this was drawn to scale as opposed to drawn to for you to  
[00:54:27] understand we don't know but it's we think it's kind of I don't know we learned something about when do these purple rectangles show up in the first place and that way and that's it thank  

## Q&A (00:54:40 – 01:13:59)

[00:54:40] you everyone and I guess yeah we'll open up the questions I think all right great Thank You Ricardo so let's see here what we're gonna do for questions is we have a number of questions in the question  
[00:54:54] and answer if you have any more questions you can submit them to the question and answer or if you would like to ask a question live with voice you can raise your hand right now I think  
[00:55:07] the best bet to start would be to take a couple of these questions in the question and an answer chat so I'll hand this over to Salim who can read the question and then you two can answer this so let's do two of those questions  
[00:55:21] and then we can try to take some sighs okay so thanks very much I just read about Macaulay want to jump in versus five so the first is could you clarify  
[00:55:31] the social cost of entry point for Aziz to jump start as currency is there any the question sorry I didn't I didn't missed you there for a second so I'm something my connection the question is  
[00:55:45] could you clarify the social cost of Sentra by a policy to jump sighs currency if there is I mean okay okay so so the amounts are this one but you should feel free again next one you should answer I'll answer outside with answering this  
[00:55:58] one um and so the answer is I'm not with it not strictly within the model but we could easily by extending the model a little bit or at least by putting the model with within a model of the  
[00:56:12] markets itself and where the exchange rates and the interest rates are not exogenous for neither endogenous and would write in but let me discuss what they would look like we haven't done it it's perfectly doable starting from what  
[00:56:24] we've done let me tell you what I what I think are the key considerations key considerations here is that all right let me start with actually an example of something else forget about the swap punch for a sec imagine that all the PBS all the Chinese let me say that the  
[00:56:38] Chinese government did was subsidize RMB credit for each RMB that any one for each loan that a one borrows in RMB from a Chinese bank or somewhere else the Chinese would the Chinese authorities  
[00:56:51] would give you 10% or something a direct credits us okay now if you're a direct credit subsidy that works in our model that's a beautiful thing meaning you can study it exactly it leads to exactly the kind of same things that I've shown you right I mean you jump-start your  
[00:57:06] currency potential insofar as you now the the distribution shifted literally horizontally literally sideways if you want because all we've done is shift the whole a bit to the left as opposed to the truncating the white tail we still have that effect now if you've done that though that  
[00:57:20] would be effective you would jump-started you get similar effects to once I discussed but potentials would be unbelievably expensive right if you think about all the credit that comes in RMB you're now having to subsidize all of it you're just it would be an enormous  
[00:57:33] amount of subsidies that the Chinese guy from would have to pay to all of these companies all over the world if you look at the following of trade flows it would really be inconceivable right so that clearly it seems that if once you add a little bit of risk or even I mean it  
[00:57:46] seems that obviously you wouldn't want to do that right or at least that it doesn't seem hard to say that that policy would be wrong with the swap lines it's more interesting right because what you're doing is cutting the right tail risk so in many ways it's so far as you've got the right only on  
[00:58:00] certain distribution it could be that it never if you never get a shot that push you to the right tail in a small sample let's say the Chinese would this would appear like it's a free lunch I mean you get essentially the same jump starting and you never the PBOC actually never  
[00:58:15] makes never lends out a single RMB that would start there and would appear like if you lunch they pretty much mean lunch does you're a small sample you're only you've only had realizations within never hitting the right tail and so all you've done is remove perceived uncertainty now I miss it you have some  
[00:58:29] realization is going to the right well then it depends on how aggressive you were in cutting that tail like Benjamin strong super aggressive in the fan people you'll see not so aggressive actually that hasn't been so used and  
[00:58:41] it's an expensive swap line still quite expensive actually um and so as a result you've only got some ideal so that's gonna cost you some then how much does it cost you well it costs essentially the interval given as you reach the expectation that you ever crossed that  
[00:58:56] tail that's how much it's gonna cost you in some ways moreover if you lent it but you then got it back then that cost you relatively small but now say that third case you end up a lot of the times on the right tail then you truncate it out  
[00:59:09] and on top of it very often the foreign central bank doesn't pay you back the RMB then you get to start getting very close to the subsidy case in the beginning okay so just in terms of just generally how much does the country lose or not it's gonna depend you Sencha what  
[00:59:23] you want to do is put in the model then the G distribution is gonna be endogenous how often you end up in the right tail is going to be something you're not just consistently potentially measure but also is it cells gonna be affected by the ball how successful the policies or not and then you could go  
[00:59:37] and answer that question that's okay so the next question at a trading firm how do you access the R or D credit does this require working for  
[00:59:51] branches of R or D banks or the all banks have access to currencies through foreign exchange markets I mean let me just say a few things about this and the code of a little tip and afterwards so in the model I guess we kind of abstract  
[01:00:04] from where it comes from we have this exhaustless cost of accessing our credit which is G function for some sort of financial friction on accessing track will and that drive things that the car though  
[01:00:18] talks about in terms of the actual are swap clients it's kind of interesting in a sense that you know what the PBOC does is it lends money to the former central  
[01:00:31] bank using a branch of an our bank as intermediaries as a clearing Bank that sits between the PBOC and the country  
[01:00:41] bank in the country to etc and so people who see credits that banks account with the RMB and that final NZO and beyond to the central bank the board and then the central atom lends did that RMB through  
[01:00:55] the swap line to its own domestic banks you can turn and it's Jim Lee to their firms the purpose of trade credit etc that's what how it works in practice or their morale is quite startled eyes and  
[01:01:07] admit some of those features okay Ricardo - like move on to next question we all have a question what's the impact of exchange rate regimes so that's sort of  
[01:01:22] discussed in opposition where is it are you going back to partition yeah yeah I guess be discussing so that lemma one be discussing how variation exchange  
[01:01:35] rates exchange why'd you want us to sue the always in your cursor and you answer the card of its value answer no I'm just playing to look at you yeah sure um yeah so Part B and discussing the role of exchange availability so obviously a  
[01:01:50] more stable exchange rate RMB side of the our side here helps the use of of the B from the point of view of why isn't that currency up and that's  
[01:02:03] actually part of the policies which are listed beginning so inside whatever it is one or two we discussed the the peg and having a having a stable exchange rate helps comments in the use it used to me prison stop stabilizers marginal cost right us that's the idea behind it  
[01:02:18] yeah that's fine so let's scroll down to the question you're switching to answer the next one selim just go sexual breath switching to an infinite horizon DSG essentially what  
[01:02:33] you'd get is right now it's kind of you go from you know here to here right what you do is you know if you had then price stickiness then last several periods some staggered way let's say Calvo or something  
[01:02:48] and likewise therefore staggering of when you call so you could have been a similar kind of infrequent updating of when you borrow when you're recovered so what you have is that instead this rectangle will grow over time right instead of going from the horizontal  
[01:03:03] where it didn't exist to here you'd have a gradual expansion of the silver time right so you have this affects you essentially get the dynamics now we initially actually even wrote some of  
[01:03:14] that because we were thinking oh maybe we should be trying to explain this effect meaning how much is the persistence in going on but then it seemed like you know that yes you would take too many pages just to explain why  
[01:03:27] this happens gradually but I think it's consistent with this happening gradually here the one thing we did empirically which is along those lines is precisely this thing I said of the process of the effect first twelve months is after alt months and she this kind of increasing  
[01:03:41] effect between the picture which would be completely consistent with having two dynamic model since you would have you'd have this thing whether it be concave or convex that again we'd have to solve the model to know for sure but you'd have to  
[01:03:53] Scrabble rice you're an excellent okay fine most question in your model if a firm uses Chinese inputs it needs to borrow RMB for button capital but the  
[01:04:07] firm has the option of borrowing USD in by OMB in the effect spot market seems the twisted input and Twitter Borel and currency can be so pumped from the firm it can be and there's a result in the paper it's not here we go it's like this  
[01:04:21] one here that requires trashed up the slide and I think actually discussed it in the presentation basically there's a very very strong force that means that the firm really wants to align the cost  
[01:04:33] of the policy of borrowing and the the currency the input is still nice to dip and it's actually pretty easy to show they want to do that and once the mixture has the same currency Boreum and  
[01:04:46] the input that service offers all that comes through a petition here so there was one I'm so similar sure I think we have one question we can ask  
[01:04:58] Fadi Hasan to ask why a new fatty here okay for a year unmuted if you'd like to ask your question sure Thank You Robin Thank You Ricardo and studying for the  
[01:05:12] very nice presentation in paper the question is is more on the empiric so in the data there was a strong decline in the share of RMB payments in 2015 but at the same time you know the deals the  
[01:05:26] swap lines were increasing so the question is was this related to the slowdown of China and can you somehow use it understand better you know the interaction between the real side and the financial side of your mechanism which i think is quite  
[01:05:39] interesting so that's that's a question basically yes right I just put here the picture just funny it's exact what you're saying right it's the follow the blue line even though red keeps on increasing just just gonna say that um  
[01:05:51] let me I'll do one very brief answer and then Saleem will touch it which is note that Assad's regime very few swap lines have been signed right so our force for blue to increase is gone now of course if everything was going on if that was the only variation that's relevant of  
[01:06:06] data our model I'd say would predict that blue should be flat not rising right because there's only two swap lines got signed here after 2015 the bulk of it came before now why does blue fall instead of being in being sane  
[01:06:19] empirically I don't know because it's in my time fixed effects obviously right so the regressions themselves don't inform on that don't tell me on that um I think you know there's two bits here so let me conjecture therefore on two things first  
[01:06:33] is like you said trade with China the slowdown of China's economy may have contributed a little bit again that wouldn't necessarily explain the fall as opposed to widen it up keep on rise right so if you have slowdown in trade and no new swap lines and no other measure not some swap lines there since  
[01:06:48] you've been no relevant measure of internationalization of the RMB in a sense of our G distributions after 2016 so nothing pushing that up trade isn't pushing it up so but now a second thing which I think is relevant and that's actually very important so far as I  
[01:07:03] didn't discuss enough in the empirics is that you know this is a date the IMF conclude the RMB in the SDR basket so there's a very important or like pauses which is the nominal sub-zero effect is actually important  
[01:07:15] for our work why I think you know and we taught to I guess I shouldn't say name so many people as I was starting this project was chatting with a lot of people me and Celine especially on the policy angle on do you think the strap-ons what do they do and we have this very strong I apologize  
[01:07:29] oh they did nothing essentially the swap lines were nothing but some posturing for including for the inclusion in the SDR once the China once the RMB got included in the SDR basket don't stop saying so many swap lines there was a  
[01:07:44] political goal that impetus was gone and therefore the RMB then was stated that the if you wanted the people you'll see ultimately and all the measures weren't really about the RMB usage so they were just about getting into the SDR basket  
[01:07:59] and that's and the army views was just an intermediate way to get there so our regressions right reject that super strongly right that would be a little iPods where our coefficient beta would be zero it seems that the swap lines increase the RB there weren't just  
[01:08:14] political posturing it wasn't just about the SDR but when you're talking about why this fall when swap lines are in there that effect of trying again the SDR that's then gone and that's all the other potential falls that are being used they stop being used and then maybe  
[01:08:27] also part of myself Salim you don't have something let me say one thing about this this picture which is more it some piece of work better on is that actually a lot of the variation in in the cross  
[01:08:42] like the total arm be share is driven by the countries that are financial centers like the UK for instance we it's a much bigger part of liberal payments and is as a part of say global economy and same  
[01:08:54] for the US for instance and so far the fluctuation you're seeing here is what's happening in in those countries because they really drive the total amount of payments that are happening across the whole world just like a way staff which is  
[01:09:09] you look as simple averages is not it's not the same you habit of same sort of flattening and maybe we should actually turn up that picture instead this is thinking about now then what you said I'm a mess that's it that's something we  
[01:09:21] should work on do not know we have we have about four more minutes I just want everyone to know that if you asked a question were not necessary gonna be able to get to all the questions but  
[01:09:34] they will be given to the authors the next question will come from Matteo Maggiore actually so Matteo you should be good to go yeah recover the first doll and Selene great paper and thank you for agreeing to be the inaugural  
[01:09:48] speaker but there was very interesting at the topic it doesn't receive nearly as much economic analysis and I wanted to understand a little bit the mechanics of swaps and supposing I live in an extreme world where then I cannot get  
[01:10:03] hold of renminbi because they have capital controls and to make a payment you need to first give me some of your currency in that sense it would be true that giving me a swap line it's a way to  
[01:10:15] get going but it's almost from mechanical there was no other way for me to make a payment if you didn't give me some of your currency first is that right is that wrong how do you think about that I'll give a very short answer  
[01:10:30] and then I think it Salim has more to say but but note material that the way the software works is it's not that Pakistan say signs a swap line you can go to the and can just say give me lots of RMB China or at last they can but  
[01:10:43] that's not the way it's mostly used the which was to use is there's a guy in Pakistan a firm whose already borrowing RMB they go in discount that load of the Pakistan central bank and then Pakistan  
[01:10:56] goes and gets the RMB okay so there was already a lending of RMB to start with so I'm sore from track let me right away put a prints that's the way it's supposed to work that's the way that again when we talk officials they would like it to work that's the way the  
[01:11:10] contract is signed having said that let me immediately open the parentheses to say that's not what Argentina did went far a lot of Rd from the from from the PBOC using the salt like just to get just to get some R&B in  
[01:11:24] order to defend its bag with a dollar but at least in theory as well as in for most of the practice material you've already you're discounting an existing R&B loan when you use the swap  
[01:11:37] line so those are NRB create to start with saline no more let me say this operational that's what you were asking how'd you get the money in in the first place so the swap line doesn't open up a  
[01:11:51] payment system right it still goes through the same structure that exists already to meet our me payments it relies on a clearing Bank typically in Hong Kong and so that that team that  
[01:12:05] must exist and it must be willing to give the central bank an R&B deposit and that's a promise it will end on DRAM B and so on and so forth and so what one does but he is just cheapens the cost of credit as it were by enabling a clearing  
[01:12:20] Bank to get R and B more cheaply from the PBOC so that's sort of the operational perspective so the actual pipes must all be in place for the money to flow now you could say that all about the swap line at all there's no chance  
[01:12:34] you could borrow it on B and therefore that's what's happening just basically just the movie a complete you know constraint anomaly borrowing all together but not a cost thing but that seemed to us a little bit too extreme I  
[01:12:48] think that was incredibly helpful I'm I think wonderful this is whether there's a clearing bag there was remember one of these controls that I told you about because again that's when you put the pipe let the thing work those controls you have a nice clear Emacs  
[01:13:02] one suggestion I had was it would be lovely to know more about when the piping essentially starts because that's another way to get the currency going is you need to lay the piping and it'd be I think you'll be interesting in in addition to being a control to just I  
[01:13:16] agree this is the thing I highlighted we have the data on whether the country in when it had an hour and be clearing back so using that itself actually the toast isn't a very interesting variable because Hong Kong does all the work for  
[01:13:30] this so I mean a bank in Nigeria can always go to its Hongkong correspondent and get the RMB so the piping is really about the Hong Kong market and that was opened in 2009 the CNH market basically and that's what  
[01:13:44] that's what connoisseur said well since that's not the key pipe these are and be clearing banks that they I think they are simple if actually the one in London is an empty empty shell so yeah the Hong  
[01:13:56] Kong market really matters