 Okay, very good. Welcome Pia, thanks everyone for bearing with us, you're welcome to to go straight to the podium if you'd like to start with your presentation, we have 25 minutes please. So sorry for the delay everyone, thanks to the organizers for inviting me, I'm actually while coming here by train I realized that I had been an intern 10 years ago at the ECB and I was driving by here and it was still a working process, so happy to be back in another rollout this time. Let's jump straight into the presentation. Slides will come in a second there. Okay, great. So what I'm going to present to you today is actually a co-author work with Matthias Kaldorf who sits in the audience, so any questions can also be I guess answered by him. We will talk about real effects of financial market integration evidence from an ECB collateral framework change. And as an introduction basically the idea what we have behind this collateral usage in a very normal and simple way, so basically you have banks which pledge collateral against if they want to participate in main refinancing operations of the ECB and this is the type of collateral that we are looking at just to make a clear distinction of what was discussed before. Somehow it's a more like a simpler role of collateral. And so the idea behind this is that we have a financial market union actually in a monetary union if there is no fiscal union then you would actually need financial market integration. Why? Because like this you basically allow that private sector funding conditions are independent of the local banking system. However, out there there's really little knowledge about the microeconomics behind a financial integration and so we zoom in into one aspect. We take you back a very long time ago in 2007 and that was a policy change at the ECB which implemented a harmonized collateral policy. And we use this framework change as a quasi-natural experiment. So to give you a short overview of what policy change we are interested in, the ECB basically implements monetary policy through national central banks. And before 2007, so you had the euro obviously but you didn't have unified collateral system. So basically before 2007 you had a two-tire system where in the first tire you had assets for example government bonds which were always eligible so they were sort of the same throughout the euro area. In the tire two you actually had assets that were under the full discretion of the national central bank. And so this on one hand accounts for the peculiarities of the local banking sector. So for example Germany is a good example for this but at the same time it also is problematic because it segments the market. And so depending on as a bank where you were headquartered you could only pledge certain types of of collaterals. And the asset we are really interested in are actually bank loans. So before the 2007 change over he only had domestic bank loans that were accepted by some national central banks. In others you really could not pledge domestic bank loans. So this can also be interpreted as an additional source of home buyers. And it actually coming back to what I was saying before it sort of violates the no sudden stop condition that you have in a financial market union. And so after 2007 what happens the ECB actually unifies the collateral framework and introduces a single list of collateral. And the consequence is that for example German banks can now pledge loans granted to Spanish firms. So in fact bank loans were one of the biggest novelties in this in this single list. Because as of now you could pledge domestic bank loans and especially you could pledge cross border bank loans. So if the borrower that you were giving money to was inside the your area you could pledge it before it was only possible domestically. And so this actually relaxes a funding constraint for banks. And at the same time it also increases the pool for firms of banks that they have access to. And so in this paper we look at how the single list affected banks credit supply. We look at how it affected cross border banking cross border lending and then we look at the real effects in the non financial sector. And the preview of the results is that how did it impact credit supply. So we use a different setup and banks holding eligible assets. These are basically our affected banks. They increase their lending by 10.6 percent compared to banks without such eligible assets after the framework change. Which firms experience the most inflows. So what we can show you is that it's mostly eligible borrows that were previously in the collateral pool that experienced those inflows. And the real effects on the firm level we have that we find effects both on the employment and on the investment side. For the related literature I'm going to make it short. But basically we are contributing to two strands of literature. The first one is cross border credit flows. So here we have in mind Janetti and Leven for example which look at how local funding conditions actually influence the so-called home the so-called flight abroad or flight home effect. And we also look at the European integration here Hoffman and Surison for example. And however what we contribute to this literature is that there is a small effect of this harmonized cross collateral policy on cross border lending because we do find that banks not only lent to their of the pool of borrows they had before but they do extend cross border lending. And the second strand of literature we contribute to is actually the bank lending channel and collateral policy. So there are other papers which have looked at collateral policy already before. One for example from Beckengabar and Irani they looked at the residential backed assets and bank lending in the Dutch market for example. And the third strand is about real effects and here Pelitsson who's who's discussed and actually has a has a paper on how the corporate bond purchases sort of affect the findings and the findings in decisions of firms. And so what we want to show is that collateral policy affects banks and you can see this by an increase in employment and investment. So I have already talked a bit about the institutional framework. I'll just now go to the data and so we look at syndicated loan market. Why? Because syndicated loan market is actually quite widely used when you look at cross border banking flows. The idea is here that you have multiple banks that lend to one borrower and we look at syndicated loan also because syndicated loans were part of being included in that list. So actually as a bank you could pledge syndicated loans afterwards. The institutional framework is also that the change the very nice thing about this change I'm trying to convince you also about this is that it's took effect in January 2007. And so it should be a crisis unrelated. We also have a very small window. We also look just at four quarters before and four quarters after. So we really try to not have the financial crisis as a confounding factor. The borrower are non-financial firms where 54 percent of them are actually headquartered in the euro area. And we merge data set in a standard procedure from Dealscan Capital IQ and CompuStat. The sample period is from 2003 to 2008. We have around 1,700 firms and we have that banks lend two-thirds domestically and one-third to other euro area banks. So the empirical strategy it's a diff and diff and our treatment and control group are basically defined by their issuance history. Because the idea behind here is that banks which already issue in the cross-border so banks which already issue to other euro area borrowers so excluding the domestic part they are actually they should be the most affected by this framework change and so we construct our affected measure by looking at the issuance history of banks giving out credits to other euro non-domestic borrowers over all credit that they give out. And the issuance history is from 2003 to 2005 Q2 and we then do a median split so basically we define our banks to be affected if they have an above median issuance history. The identifying assumption here is really that unaffected banks do not change a business model in response to the collateral framework change and we estimate both in terms of loan issuance and in terms of interest rate spreads we estimate the following equation. So it's basically our dependent variable of interest so a loan volume or interest rate times our treatment variable times post and we include to alleviate some concerns we include bank level controls so the concern might be that there is no random treatment assignment and we also include bank firm fixed effects to account for different borrow characteristics and we also include firm quarter fixed effects in the spirit of Kwajamiyan so we take out the loan demand. We also include it's not on this slide anymore but we also include a very stringent industry country fixed effects in the spirit of the Greiser 2019. So to show you the parallel trends now parallel trends basically is to underline the validity of our diff and diff and here you can see the loan volume and as you can see there is a little bump going upwards but it's actually not significant so you can still argue that the issuance behavior of both the affected and the unaffected banks is the same before the framework change happened after the framework change you see that both start diverging. The same is what you can see in terms of interest rates again here there is no the difference between our treatment and control group is null basically before the framework change and after the framework change you see a dip in the interest rate spread. So our main table is basically the following here we have the credit supply as a dependent variable and we interact our affected measure with the post framework change measure and what you can see is that we have a highly significant coefficient throughout all our models we basically just add more stringent fixed effects going from column one to column two to column three and so in column two actually you have the model with a firm time fixed effects and in column three you have country times industry times time fixed effects and as you can see the coefficient does not change that much and we find that banks which are affected by the framework change increase their lending by around 10% compared to banks which were not affected by the framework change. We also include a list of bank level and loan level controls which didn't have space on this slide so I just wrote it but the interest rate spread is basically here we just change the dependent variable and we look at the spread of the loan above the LIBOR basically that's the reference rate in that market and what we find is that given that banks have a positive funding shock right we the first hypothesis is that through this positive funding shock they increase the loan supply and at the same time they should decrease the interest rate spread right and so this is actually what we find as well in terms of basis points we find that banks reduce the interest rate spread by 12 basis points. Looking at the borrow location we were interested in where the borrower is located and here we basically make a sample split so it's the same specification than before but we just define borrowers if they were already previously eligible or if they are newly eligible and what we can see is that banks sort of used this framework change to give out more money to the borrowers that were already previously eligible but also a bit to borrowers that were newly eligible why do I say just a bit because if you look at the magnitude it's basically one-tenth if you look at the newly eligible and the newly eligible is mainly the non-domestic euro area borrowers and so this is sort of our hint at how this collateral framework change actually increased cross-border lending. We then go to the how am I with the time is it? We have eight minutes. Okay great that's perfect okay we basically do the same at the firm level so we look at the same coin but just from the firm level side and here we define exposed so a firm being exposed to an affected bank is basically we look at the amount of loan that a firm gets from affected banks okay so we again look at the time period between 2003 and 2005 and we have loans from affected banks over all loans and again we do a median split so firms with an above median share of loans coming from those affected banks actually they are classified as exposed okay and here we basically have again loan issuance so the probability of in this time it's the probability of a firm getting a loan times beta one here the coefficient of interest is actually the interaction between our exposed measure and the post again we include this time firm controls and we work with industry time fixed effects and country time fixed effects as well as firm fixed effects and what we find is that we have some credit supply some positive credit supply and we also have real effects in terms of employment and tangible assets and so in the first slide what you can see is basically the other side of the coin of what I showed you on the bank firm level this is on the firm level again we see a positive impact so firms with a relationship to affected banks actually have an increase in the probability of increase in the probability of getting a loan same in terms of loan volume they experience a positive increase in loan volume and for the real effects actually here we this is still a bit in the making but here we basically want to know who are the borrowers that experience most of the benefits and we we try to split the borrowers into the tradable non-tradable sector as a proxy as well risky or non-risky and so here what I show you is the tradable non-tradable and what we find that both in terms of employment and tangible assets actually if we split the sample into those we find also some real effects so it's mostly firms in the tradable sectors which with a relationship to affected banks which experience an increase both in terms of employment and in terms of tangible assets we also did a bunch of robustness checks I don't think I have time to go through all of them but I just want to show you two or three which I think should should sort of convince you about our story so the first one is when there was this collateral change actually the the idea that we had is that we would only look at something that was a novelty for all banks and that was that they could now pledge cross-border loans however for countries like Italy they actually could not pledge not even domestic loans before this collateral change and so it could be that you know once Italian banks could pledge domestic loans that it is really that which is driving our results and not the cross-border pledge and so what we do here is that we just construct our affected measure again including also the part where banks can pledge the domestic part newly so that's basically an Italian fixed effects so to say and what you can see is still that our results hold and are quite constant we then do a triple interaction of this just to see what drives what and here the results disappear which means that it's not the domestic part which is explaining our results it's really this euro area cross-border pledge eligibility and the possibility of pledging cross-border loans which drives our results one other thing I also would like to show you is a placebo test so what we did is that we constructed a fake affected measure for banks not part of the euro area so mostly actually it's 2007 so it's mostly UK banks still in the EU it's we just look at European Union banks not being part of the euro area and we construct the same measure and the idea is that ECB monetary policy especially a framework change should impact only euro area banks right and so here we do not find results and that's actually good because it means that we really do find a peculiar thing just for euro area banks and we also change a bit the affected measures we define it over total assets we also look at the continuous treatment variable as a share and we do find always consistent positive and significant results as a last check we look at the whole sample where we don't look at term loans but sort of revolving loans and things like this and we also don't find results so let me conclude we look at harmonizing collateral policy as one aspect of banking union and our idea behind this is that an increase in collateral actually leads to a positive funding shock for banks and so we should find positive results on the loan level and this is what we what we find we find that increases bank lending and it increases bank lending also cross border our results also suggest some positive real effects what are the downside of being a single list well cross border capital could fuel unsustainable credit booms this is what we have seen in the run-up of the financial crisis but we don't have any counterfactual this is a different and the limitations of our analysis and I want to be very clear about this is that local funding conditions still diverged after 2008 so a collateral policy definitely was not the silver bullet to sort of resolve the sudden stop condition in the euro area thanks a lot we would like to take a seat so to discuss the paper we have professor loriana peloton from the gethic university frankfurt who will be joining remotely and i hope loriana you can you can hear us and we can hear you i can perfectly hear you i'm just trying to share my slide so good good morning to everybody i hope that you are able to see my slides it's fine not yet but we're getting there but we can hear you so that's a first step good so thank you very much for inviting me to discuss the paper are you hearing me everything is fine yes we can hear you well loriana but just give us a second to switch the screen here so we can we can see the slides yes yes so you know it is a very interesting paper it has been presented very well and the only concern that i have is that really the version that i have of the paper is not the last one so part of my comments maybe are not really related to what you just saw about the paper but anyway uh what is the paper trying to do is to the main focus at least this is what is claimed in the title is to estimate the real effects of the armonite collateral policy that the ecd has been introduced in southern south and they are trying to measure real effects by looking to employment and the the key focus that it was at least in the in the version of the paper that they read it was on weather is changing the collateral list create any cross border credit provision so any fact across both borders and as you in some sense observed by the presentation but it has changed a little bit in terms of of the results it seems that indeed this list changing the list has been improved the banking union if you want so core countries banks uh lent more to peripheral countries reducing their own bias and this was pretty much the result in the version that i read of the paper they do send me the slides but it was yesterday afternoon and you know it was very difficult to understand what was there um so we're coming out by this change of the list is that the credit supplies has been increased uh largely to uh riskier and less productive firms in periphery countries but this has also generated an increase in employment and investments and as you observe in the result the cost of credit is being reduced by 12 by this point so uh i think that the topic is extremely important and it is uh really helping us to understand how important are these eligible lists set up by the ecd so clearly the topic is very interesting and i'm working a lot on this topic on the corporate bond side on the sovereign bond side and i think that this paper is very complementary to this type of literature because it's really focusing on uh the the lending part that's so far has not been investigated and uh but but my point my first point is that uh the main purpose of of the ecd for creating a unique list was to improve monetary policy transmission and i wonder if really this paper is able to answer to this question you know it's trying to see if we have any real effect but it would be nice also to figure out if also monetary policy transmission has been improved thanks to this change of the legibility list moving from the fragmented list at the national level into a unique list at the ecd level then there is another aspect that i think it is important to to to stress uh the the change in the legibility list happens on january 2007 but uh even if the global financial crisis is associated with let's say the the liman default that was in september 2008 so let's say two years after this change or let's say one year and one half after the change of this legible list actually the crisis started in august 2007 uh i don't know you know it's funny because when i'm asking to my students when the global financial crisis started most of them are answering september 2008 because they are related to uh they relate it to the liman uh the fold but actually the main crisis started on the eighth nine of august 2007 when bnt paribas or europe went to the ecd asking money because he was not able to raise any money in the enterprise market at that time uh and he needs this money mostly to provide funding to his co-connected edge funds that was largely exposed on acid-baked corporate uh commercial papers so and you know the proof is this graph this is reporting the three month library spread and as you can see there was a huge increase of this spread by 100 basis point in pretty much one day exactly in uh august 2007 so why i'm talking about this well because you know when you're looking to the difference and the improvement uh you're looking to even when you are shorting your uh your windows you're looking to one year before and one year after till the fourth quarter of 2007 but the fourth quarter of 2007 is including already two quarters that are subject to the crisis and in fact if you're looking to your results well you can see that maybe there is something in the first two quarters but already at the third and the fourth quarters the results are not anymore there so you know you have that pretty much the increasing landing and the impact on the spread shade away uh starting exactly from the third quarter of 2007 so i wonder if uh you know the result that you are showing to us is just driven by the first two quarters and then clearly the crisis prevents you to to have any uh let's say uh possibility to do any empirical analysis on the implication of this changing the in the eligibility list now i have several other questions regarding the paper the first is that you know since all the story is based on the fact that by changing the eligibility list banks are now able to use different assets uh different loans and before uh well you should be able to answer to this question by looking indeed if banks are now pledging new loans to the ECB you know being at the at least at the Wundersbank uh you should have this type of information so it will be very interesting to see not only if they increase landing but because there are the possibilities of why the increased landing can be several but the key point is are they really using uh and how much uh you know the loans that they had that before they cannot use as collateral uh now in order to get uh new funding from the the ECB so pretty much look to the loans that before January 2007 would not be considered flexible at the ECB at least in some countries and document that indeed after January 2007 banks do pledge these loans as the ECBs you know uh very uh interesting because it will show that indeed there was an impact and how much uh then i have some question about the result that you have uh on cross-border landing uh i don't know how robust is this result but it seems that cross-border landing is increasing but largely for riskier and less productive first and you know i'm not debating if this result is true or not my question is why is this the case so uh it is because these are the only available options for no incumbent banks so you know if you had a foreign bank that start to lend more on uh uh for two firms on another countries truly at least at the beginning you will have the leftovers because you know all the story regarding relationship in landings and so on will tell you that will be very difficult for the new banks to come and pledge a way uh lending uh to firms that very good so clearly i would like to see something more about this uh uh this result and and also it will be nice to to investigate more how competition among banks has been changed due to the unique list so you know they start to uh land in places where they were not lending too much or they have the capacity now to lend more to these other areas uh how this is changing the competition you know what is going on there it will be very interesting to to investigate um then i have another question you know uh before let's say august 2007 in europe we were having a good interbank market and the interbank market at the time it was it was yes cluster a little bit at the country level but there was a lot of cross-border landing among banks that then it has been uh let's say frees up and uh and destroyed with the with the global financial crisis but uh one question that then i will be curious to know is that why the interbank market was not enough for transferring funding that in some sense was some funding surplus cross border before so what what was going on here why did the bank market was not enough to uh provide this type of uh let's say cross funding and you need to have a unique eligible list in order to improve lending across countries and uh and also how much is important for banks to try to exploit this possibility in terms of diversification of their portfolio uh is is really banks that now are able to diversify more than one that are doing exploiting these possibilities or are some others so you know what are the main drivers uh still this paper i think is not telling us uh completely the story and then going back to some of the regression you know you show that there is some statistical significance and then for example for for the lending there is an increase of lending by 10 percent with respect to before uh but how economically significant are they are this result how many billions are we lending more thanks to this uh uh let's say um changing the uh eligibility list you know and uh because on the other side from the spread 12 basis points seems important but we are talking usually about 200 basis points so you know if you're considering that there is a lot of dispassion maybe these 12 basis points are not so relevant and then this impact is not so permanent it is true that it's due to the crisis that we have after but i wonder if there was just a uh a short term effect or if we are continuing or not to see this type of converter effect thanks to this eligibility and also regarding the riskiness of of the firm you're looking to the return on asset but i think that you should think to some other measure for this type of uh uh let's say indicators and finally the title seems you know to induce the reader to expect a lot of analysis on the real effect of the change in eligibility list but actually this result is arriving only on the end and in fact also in your presentation you just present one slide regarding to this result so i wonder if really the focus of your paper is on really the real effect or simply on the increase in credit and lending uh due to this eligibility list or you know uh something else so either the title will be more broad or the focus actually really just on uh or mostly on the real effects of your analysis and your analysis for the change of the legibility but in any case very interesting paper i'm learning a lot and i think that is contributing significantly to the literature because it's really looking to one part of this uh let's say of the change in the legibility list induced by the ecb that has not been investigated before so thank you very much so thank you very much loriana um a lot of questions so uh i'm sure you appear you're keen to get back to some of these points but if you allow we take one or two if there are questions also from the audience and then we give the floor back to you for a collective response so any any questions john david and francisco so can we have the mic please yes just a quick question um so on your your presentation you talked about affected and non-affected entities i think there's a there's a probability that there is no such thing as non-affected entities in the sense that if you think that there's a market for such assets and there's a competition for this for this asset the affected are now have an advantage to hold these assets and the non-affected have a disadvantage they don't have this advantage and they are losing out and why am i saying this is that um you may double counting uh you may be double counting the effect in term of volume in the sense that the affected will increase their holding of these assets and what you call the non-affected will decrease the holding of these assets and your diff and diff will double count uh the effect i think economically it's probably fine but when you discuss the magnitude especially in term of volume i would be careful um about this this potential effect thank you thank you john david uh francisco great presentation once you're done with all the comments of loriana going back to one of your early slides that you contribute to several transfer literature could you be also looking at the impact on resharing within the you especially the euro area the the no-con effect on the fdis uh intra euro area fdis which were very high and but especially resharing thanks okay thank you francisco i think we we conclude the collection of questions now because we're also running out of time so we can give pia the chance to respond yes okay so um thanks loriana for for your presentation you are actually the the best uh discussion we could have hoped for because we we obviously know your work about uh quadraturel as well um and i do apologize for sending a new version only yesterday but i have been working quite extensively on this project and um yes so thanks nevertheless for for your very helpful comments um um we always get this comment that uh there is differential crisis and that it it is a confounding factor and we have to take care of it so what we normally do is that we shorten the event horizon just to the end of 2007 but i'm totally aware that uh the crisis started before there's also papers out there by roholl and puri showing that german banks already had difficulties in in in um the autumn of 2007 so um indeed what we argue is that it's it's given that we assume a positive shock to bank funding we find a positive um credit supply this sort of confounding factor might at most bias our results downwards and and this is a bit what you have argued right and so if we cannot convince you that the financial crisis uh happened in 2008 uh then uh indeed at most it biases our results a bit downward that's uh that that we have to admit indeed um about what banks pledge so i i started this project as a phd student and i did not have data on on what banks actually pledge at at the ecb uh and so we could obviously just uh just um build a proxy however we look at as as a as a proxy and we also have this in the data we look at um the increase in non-marketable assets which are mainly bank loans uh before and after the framework change and you can really see that they doubled uh in in the ecb collateral data so this is at least even though we don't have the the bank level information it is at least a hint that you know something changed uh in in in what banks pledged and they used this however it's a very aggregated measure um yeah and on the cross border landing we will uh and also when it comes to competition uh we will we will definitely have a look in this uh point well taken and also your first point on the first order effect of um if it really improved monetary policy also this it's a point very well taken thank you um yes and the real effects we are working on on on getting more out there basically that's that's that's the that's the construction works right now they are right there thank you then i would go to the to the audience but again loriana thanks for for for your comments um be affected it's uh what we what we do is um i wouldn't talk about double counting to be honest because i would not understand why uh an unaffected would find it uh um why an unaffected banks uh would be be negatively affected by it this this this you have to explain it a bit more because if basically we distinguish our measure distinguishes between banks which are doing cross border activity in uh to other euro area countries and banks which are not really doing it so it's banks mainly focusing on on domestic activities right so our extinction is basically between more international banks and less international banks and so um that's that's the distinction we do so i i wonder where this okay yeah totally and uh also for the for the risk sharing um thanks thanks a lot for pointing out the fdi part of it uh i i was not aware but we will look into this yeah thank you frances okay thank you pia thank you loriana uh many thanks also uh to to uh participants for the questions and the interest in the paper um so we conclude our session on collateral here uh we'll take a a break a coffee break until half past 11 and we'll be back with the session on repo so um perhaps you can join me in thanking all presenters