 So, good afternoon everybody, I am privileged and honoured for me to welcome the speakers and one discussant, this is Efi Papa from the Universitat Carlos de Cerro de Madrid. Discussant is Roy Betzmer and then we have the second presentation with Marco Passetto from the Federal Reserve Bank of Minneapolis and unfortunately our discussant there could not make it, so Marco can speak a bit longer. As I just said, Efi is professor for macroeconomics at the Universitat Carlos de Cerro de Madrid and she holds also a PhD from Pompéu Papa. Her research is focusing on fiscal policy monetary economics, opening economy and macroeconomics business cycle, so this is exactly the right research focus for our conference here. And she will speak about the likely macroeconomic effects from the EU recovery plan. Well, thanks a lot and we have already set up half an hour to speak. Efi, and afterwards Roy and then afterwards the discussion, the floor is yours. Sorry, I forgot, sorry, I forgot, sorry, just one remark that all participants can raise questions via the chat function and I can then look at them and depending on how much time we have we can give them to Efi and later Marco, but you can say whether you want to have questions during your talk or only at the end. So sorry, Efi, now the floor is yours. Well, I'm honored to be a participant in this conference. We have already listened like very interesting views about fiscal policy and its interaction with monetary policy. In this talk, I will concentrate on fiscal policy issues and yes, you are free to ask me questions during the presentation. So please go ahead. So I start my presentation by giving you this enlightening picture from the unsyntagorized happens and with enlightened by the next generation in you and the slogan that we will make it real. So I thought also in the previous presentation, there was some kind of moderate enthusiasm about the effects of the NTU and the effects of the fiscal policy in this recent current crisis. So, however, you know, we don't really know what is going to be the success of this NTU fans on really bringing growth and recovery as they are supposed to. So what we do in this paper with Fabio Canova is that we are going to look at the likely macroeconomic effects of the EU recovery plan. And in order to do so, we are going to go to the past experience and try to draw inference from other funds that they look like the NTU funds and they have been actually administered by the EU in different European regions and try to evaluate what kind of macroeconomic effects these funds had in the past. So I think that everybody in the public knows what the NTU funds are about. Basically, the majority of the funds are going to come in terms of loans or in terms of grants that they are going to supposedly increase the human capital in EU economies. They will lead to renovation, scaling up digitalization and upscale and rescale activities. So the issue is will the great jobs, will their conversion to a greener economy be smooth? So all these kinds of questions that are written in this slide come about when we think about the NTU funds. So what we will do in this paper is we will try to study the regional dynamics produced by major past EU funds that they share characteristics with NTU funds. In particular, we are going to look at two funds, the European Regional Development Fund, which was launched to foster R&D innovation and to favor the digital agenda and to support small and medium enterprises. And we will also look to the European Social Fund, which was launched to support investments in education and health and also to provide poverty. We will construct a novel database of these regional funds to get and we will combine it with the macroeconomic regional data. And we are going then to try to collect some stylized facts about the macroeconomic effects of these funds and try also to highlight whenever we find them regional heterogeneities. At the end, we will present a model that is going to rationalize the average effects of the two funds and also can account for some heterogeneities and we will use the model for policy advice. I don't want to spend time telling you the results. I prefer to show them to you directly. I will not talk about the related literature. If some question comes about later on the discussion, I will do that. So let me explain you a little bit the data that we're using. Here we are going to use regional macroeconomic data that come from our Deco and they are available online and we will restrict attention to 279 knots two units. So this is something between the country unit and the very local unit. So it's like the county in the US data, let's say. So we will have annual data on real gross value audit, on employment, real compensation, population investment. And we will construct also cities for labor productivity. And we will use data on European funds, looking at the historical archives that we have on EU funds. There are various programs that the EU has actually administered as funds to the different European regions, the Cohesion Fund, the Agricultural Fund and the Fishery Fund lately. But we will concentrate on the ERDF, so the Research and Development and the ESF, the Social Fund, because we consider that these two funds are the ones that they mostly resemble the NGQ mutual funds. So the data, they're kind of difficult to handle. We had to make them real. We had to adjust for gaps. There are some issues about how the expenditures are actually measured in the books. We try to arrange all these kind of problems that we had in the data. And with this novel data set, we are going to go to look at the question of interest, which is what is the macroeconomic effects of these funds. Just to give you an idea about how these funds looks like, this is the distribution of all the regional funds in the different European regions. So as you can see, there are some recipients, like, for example, Spain and Greece and the south of Italy, as well as the Assesson countries, that they receive much more funds per capita, relative like to the central Europe, like France and Germany. Now, what we're going to do is that we will take your base and approach because we have only 30 years of data. And in this sense, what we will do is that we will regress the variable of interest that for us, it could be the GVA growth or the employment growth or investment growth, et cetera. And we will regress it on its own lags. And at the same time, we want to introduce this fund source. So the way we are going to create this fund source is by cleaning the funds from variations that they have to do with the European cycle. So we will do a two-stage regressions. In the first stage, we are going to take the funds and we will regress them against like a constant. And you are a variable such as your GDP employment, the deflator, the nominal exchange rate. And the residuals of these regressions, we are going to use them as instruments in the regression that you see in the first line of the slide. So actually, the way we are going to perform this regression, YIT is going to be the cumulative effect between period T and period H, we are going to use local projections here of the variable of interest, let's say GVA, while XIT H is going to measure the change in the fund that we are going to actually normalize with the GVA of the region. Why are we doing that? Well, because in this way, the coefficient C I H will have a natural interpretation, which is going to be the multiplier of the fund. Now, because we have some issues with dynamic heterogenetics in the sample, what we're going to do is that we are going to run these local projections region by region for the 279 regions. And we are going to construct a distribution of multipliers that we will then try to cluster according to different characteristics, as you will see very soon. So as I said, because we have a short sample, we are going to use Bayesian approach and we are going to impose some priors for the coefficients of interest and for the various matrix of error and we are trying to be as agnostic as possible as possible about these priors. And we will also try to control for anticipation and as well as we are going to construct cross sectional distributions of multipliers, as I told you, we're going to have a distribution of 279 multipliers that we're going to discipline and the way we're going to discipline it is by clustering the units according to different characteristics of the regions. And what we will report on the benchmark results is going to be trimmed average effects. So let me go directly to the results. Somebody makes too much noise down here. So what you see here is the average cumulative effects of the multipliers for the different variables of interest. So here on the left column you see the variables of interest that we consider. So we have GVA, employment, compensation, investments, labor productivity and labor force participation. And we have the multipliers for one, two and three years, these multipliers are cumulative for the ERDF and the ESF funds. So as you can see, the first thing to notice is that ERDF funds, if we look at GVA for example, have an immediate positive multiplier effect, which is pretty big. We are talking about a multiplier of 1.83. Usually the multipliers that we get when we look at government consumption are below one. So getting a multiplier of 1.83 is significant and important economically multiplier at the first year horizon. However, as the horizon goes by, these cumulative multipliers, we see that this multipliers decrease, which means that the effect of the shock dissipates as the time goes by. And we see this kind of pattern also if we look at other macro economic variables like, for example, investment. We see investment has a similar pattern. And the same is true also for compensation and labor productivity. The only variable in this multipliers for which it looks like that the effect of the ERDF funds persist is participation. When we move now to look at the ESF funds, we see a very different picture. Here multipliers on impact are insignificant and if anything negative, but as time goes by, their effect cumulates, which means that they continue to enhance the economy in the medium run. And we see that if we look at the multipliers for GVA, but we can also see this a similar picture if we look at investments or also a labor productivity and also like in participation. So it appears that if we look at the effects of these funds, it looks like that funds that they are going to promote, the research and development, the digital agenda and small and medium enterprises are going to have effects that they are going to be significant, but they are going to be short lived. And instead if we look at funds that they look at health and education, we are going to have insignificant impact effects, but we will have very strong medium impact effects. So now we are going to do in the analysis a lot of sensitivity analysis. We have 279 units there, the regions for which we include also the UK. So what happens if we exclude the UK or we look only at euro area members or we look at the data after 2000. This is what we do in this graph and for comparisons, we also report here the baseline results. So what we observe basically is that the ERTF funds multipliers are kind of sensitive to what kind of sample we are going to use. So for example, if we use like only euro area countries, it appears, I'm really sorry for the noise, it appears that the multiplier effects are stronger for your countries, even three years after the shock. And it is not the case, for example, if we look at at the countries that include UK. And with respect to the ESF funds multipliers, we have a robust picture in the sense that in all the experiments that we do, we also observe, we always observe the same pattern, which is zero impact effects, but very strong cumulative effects after three years. Now, we have also looked at spillover effects and we wanted to investigate whether the effects of the multipliers are stronger in the region or they actually spillover to the whole country. So here what we are calculating is cumulative multipliers at the national level of the same funds. And if these numbers are higher than what we have seen at the regional level, it means that we have positive spillover effects. If they're smaller, the spillovers are actually negative. So what we see is for the ERDF funds, we seem to have some positive spillovers, but they are not really significant. And if we look at the ESF funds, what we observe is that actually we have some negative spillovers. And these negative spillovers come from actually the investments crowding out. If you look at the fourth row of this table, what you would observe is that investment multipliers at the national level for these funds are actually negative, which implies that if I have an investment in Andalusia in Spain, this is going to crowd out investments in other regions in Spain like Madrid, Barcelona or Asturias. So to give you an idea of what these negatives spillover would be. So then we have looked at heterogeneous effects. So we have compared northern regions versus southern regions and all regions of the EU versus younger regions of the EU. So what we observe here is that it matters where you are located for the size of the multiplier. And it seems like that both kinds of funds have much stronger effects if you are located in the south, where the south, the south here is like loosely determined. It is Bulgaria, Cyprus, Germany, Spain, Croatia, Hungary, Italy, Portugal, Romania and Slovakia and Slovenia. Sorry. And it also looks like that if I belong to a young member of the EU, like Bulgaria, Cyprus, Czech Republic, Estonia, Croatia, Hungary, Latvia, Poland, Romania, Slovenia or Slovakia, the multipliers dissipate extremely fast. They have, for example, for ARDF funds, they have a positive effect on input that dissipates immediately after two years. While the ESF funds, they have effects that they remain for two years, but they start dissipate in the third year after the shock, which means that here the medium run effects for these kinds of countries are much smaller. Now, another exercise that we do, which we think is very important, is to cluster our units according to the income per capita in the different regions that we consider. So we will construct four quantiles for the income distribution and we will look at how the multipliers one, two and three years of the ARDF funds. So here we look at the multipliers of all the variables that we have in the analysis, chains with the quantile of income that we consider. The takeaway of this picture is that multipliers for ARDF funds are going to be always more significant on impact, but also in the medium run, for regions that they belong to, that they belong to in the second and the third quantile. What does this mean? This means that these regions are going to take the most out of these funds and actually they are going to cut up with the REITs, while when we look at the poor and the REITs we see that the effects of these funds are the regions for which the effects of these funds is actually kind of temporary. Now, if we look at ESF funds, we will see a very similar picture. Again, we look at quartiles of the income distribution and what we look is at the one, two and three year multipliers and here I saw your GVA but you can see that for all the variables that we consider the effects are similar, we see again it is that the medium income regions that mostly take advantage of these NGU funds and they grow faster after three years, while we see that the multipliers for the first quantile dissipate pretty fast and it looks like that for these funds also the REITs kind of taking advantage of these funds also in the medium run. So this is a picture that shows that whether I'm going to look at ESF or ERTF funds it looks like that both funds might create income polarization and this is something that we should be aware of. So if we now cluster multipliers at the country level and we look for for example our ERTF funds this map is a map in which red is bad and purple is good. Why? Because red depicts multipliers in regions for GVA and employment that they are both negative after three years and the purple instead is good because we are representing with purple here the regions for which three years GVA and employment multipliers are positive after three years. So what do we observe is here we have a polarization we have regions in Europe in which we have a lot of purples and regions in Europe for which we have a lot of red which means that you know regions in which the multipliers for ERTF works and regions in which it does not work and we have some only few regions in the middle for a few countries the middle like Portugal for which for example we have negative GVA multipliers but positive employment multipliers and the green regions here are the opposite our regions in which countries in which we have positive GVA multipliers but negative employment multipliers. So it looks like that not everybody will benefit the same from this kind of funds if we look at RDF but if then we move at ESF funds we see a more purple picture and purple and green which means that this is good so now all central Europe looks like to be affected positive by the ESF funds and we have very few regions that they are still in red in which the funds do not have any medium-run effects. So let me conclude and try to recap my facts before I go to the theoretical model so what I saw you so far is that ERDF funds seems to be useful for short-run purposes while ESF funds look to be better suited to force the medium-term objectives. The macroeconomic effects of the ERDF funds are kind of less robust but in general they are not persistent while the ESF funds have more persistent effects. We have regional dynamic responses that are very heterogeneous we have country specific features like location, tenure in the EU and the level of development that matter for these differences and in the response to these programs it looks like that lower peripheral sorry a newly tenured regions of the EU do not take full advantage of these funds and underperform in terms of recovery and transformation. So having said that I would like now to go to try to interpret in the 10 minutes that I have left my the dynamics that we see in the data. So what we will do now is that we are going to build a two-country model of a monetary union where we are going to have a small country which is going to be a home country which is going to be small and it's going to be the region within like a monetary union and this kind of model we have standard new Key Nation features because we want to generate effects of these funds that they will have like that will be important for demand so we want to generate demand effects from fiscal policy but at the same time we also want to generate supply effects from fiscal policy because everybody said in the in the past one hour that what we expect from the NTU funds is growth and we want to have G increase in more than R so this means that we want to incorporate in our analysis and in contrast with previous theoretical models about fiscal policy growth and fiscal policy affecting growth for that reason we are going to have endogenous growth that is going to come through R&D and human capital accumulation and we are going to assume that the federal expenditure is going to be financed with local income and lump sum taxes and the federal government is going to provide funds that they're going to enhance R&D and human capital accumulation in the home economy the monetary policy is going to be conducted with the Taylor rule at the union level and we are going to try to look at different ways the funds might be administered in the regional economy so what is key here and I want to highlight that is that the federal government spending is going to affect the R&D and the human capital accumulation so I don't have time to explain you in detail the model but basically what I'm going to have is agents that they will derive utility from consumption and this consumption group is going to be a tradeable good and so it's going to be a good that it is domestically produced with which I write here as CHD and CFT is going to be the composite of goods that they come from abroad ETA is going to determine what is the home bias in consumption and the agents in the economy are going to derive utility from this composite consumption good and this utility from working and from going to the school or getting educated so MT here is that this utility that I get from getting educated so now what we're going to assume is that we will have a human capital accumulation equation in which HD is human capital and it's going to accumulate as almost it happens in all these models of human capital accumulation the sense it will depend on human capital in the previous period but it's going to depreciate and it might depreciate faster depending on the utilization of human capital but what is more important is that in order to be the human capital I have to study and studying together with the aggregate human capital is going to generate the new production of human capital what we're going to assume in this model is that it can be that by giving funds to human capital accumulation the government is enhancing the investment in human capital so that's what we're going to call HK accumulation we might also think that actually the NGU funds are not going to come as direct subsidies to investment in human capital but as, sorry, spending in human capital but they might come as subsidies to human capital so in this case what we will assume is that in the household passes the consent we will have this term STM in yellow in this type which is going to be the subsidy that the agents are going to receive for allocating time in education so and we will call this a case in ormo in our slides later on HK HK subsidies so we assume that the NGU funds can be either administered directly through investment in human capital accumulation or through subsidies in human capital now I will not have time to describe very much the economy in detail so I will just move in the last time minister tell you what is the R&D accumulation in the model here when we're going to to assume that we have R&D accumulation we will assume that the R&D accumulation in the economy can actually affect directly the production function and in particular it can affect the TFP in the economy the TFP is this variable ZT and we are assuming that the ZT is actually going to depend on R&D investment R&D capital and GTRD is going to affect the labor augmenting total factor productivity in the way that I present here in the second equation or if you don't like this assumption we can also assume that the government as in the case of a human capital is directly investing in the accumulation of R&D and this is going to be through the accumulation of R&D in the third equation here and as you can see we have this function here omega in which government spending in R&D is going to enter the accumulation of R&D capital or if you don't like this assumption we can also assume that the R&D is coming to wholesale firms as subsidies so the government can administer the R&D funds either directly enhancing TFP or through the accumulation of R&D or through R&D substance so here I don't have time to talk to you about the responses of the economy and what happens more or less so let me move directly to show you the multipliers that the model generates basically we are able to match what we see in the data we have open multipliers that they are very high on impact and they dissipate as times goes by especially for the case in which we assume that the R&D is the G funds go to R&D accumulation or to R&D substance and less so when we look at the TFP what is the intuition behind this result? now I will go to my slides so the intuition behind this result is very simple if I give money as fiscal spending in the accumulation of R&D and I don't give them and this money do not enhance directly TFP then what is going to happen is the increasing government spending coming to R&D investment is not going to come hand in hand with an increase in the capital investment instead if I manage to affect TFP directly with my NGU funds with my R&D investment then you see the blue line here tells you that I'm going to increase the capital investment and this is going to generate more persistent effects of R&D in the economy now if we look at ESF funds we have the pattern of responses that I show you in the empirical model we don't match the magnitudes that is because the model has its limitations but the mechanism behind this kind of multipliers is best explained here when we have an increase in human capital investments what is going to happen is that people will move out of the labor force and they will go to study so they will increase the education in hours now this education in hours is going to bring down the working hours and it's going to make labor very expensive so in the short run this is going to have cost but as the increase in hours is going to enhance the accumulation of human capital this effect is going to dominate and we are going to see the positive medium multipliers that we see in the model so I have probably one minute we have tried to explain also the intelligently I don't have time to explain you but I want to finish my presentation with is the case of grants versus loans in the data we have grants but in reality in NGU funds we have also loans so inside the model we can address the issue what happens if I use loans and grants instead of just grants to finance these NGU investments so the meaning behind this picture that probably you don't understand anything because there's so many things tagged here is that if I use loans I am going to have short run effects negative effects from the NGU funds so the multipliers are going to be much smaller and especially in some cases much smaller than what I would get if I would have only grants so I want not to run out of time so let me recap I try to show you what are the possible effects of NGU funds the outcome looks moderate moderate optimistic but we have to be careful because there can be reasons to worry about polarization in the model we suggest policies that will make the outcome more optimistic and we suggest like what kind of administration of funds as well as what kind of structural characteristics are needed for these funds to work at their maximum and we have also been able to rationalize the average effect now I would like to thank you for your attention I think I'm a little bit over time I think that we should all think that the future of Europe is going to be as these kids have depicted it in this picture bright and full of stars and I personally think that we can make it and this is also what my research outcome suggests thank you for your attention wonderful wonderful the terrific presentation and research work I'm not going to ask a question now and there have been already several questions in the chat we are going to that after the discussions but thank you Evie that was really and also the timing was spot on I took two minutes and you took 40 minutes now I'm going directly to to the room in Bezmar who all of us know is the European Prince's Board and University of Amsterdam the floor is yours yeah thank you Klaus and thanks the organization for having me let me just see can you I suppose you can see my slides right so I was quite happy with this in our final slide by Av which is on the optimistic note I'm also optimistic I think probably the future of Europe is better than many of us think I think this paper is it's a very very rich paper and it deals with very important issue and that has not been really been analyzed I mean not in the well in an econometric quantitative way in the way that Fabio and Av do and so what are the likely macro effects of the EU recovery plan and there are a number of of course a number of important questions when it comes to the RF so one of the questions is when will it be considered a success and that is a very important question because if it is a success it can be a stepping stone for a permanent fund to strengthen the structure of the EU economy or maybe a bigger EU budget and potentially pave the way for a central fiscal capacity which was also discussed in the previous session and another question is is the effect of the RF is a temporary or long term so temporary I mean it doesn't cause a shift in output or does it really cause an increase in growth and sorry I mean this is my institution tries to save energy so the lights disappear sometimes the third question what are the relative roles on expanding the supply side and increasing demand because the RF has you know affects both sides and I will say something about that and what the paper could not really address because it's you know how would you say it's very in a way not so tangible and it differs a lot across countries is the role of the of the reform component and how that interacts with the investment component of of the you know of the facility so the attribution of this paper tries to affect the assess the effects of RF as I said important interesting it really addresses a gap in the literature and importantly in this paper the assessment can only be done indirectly necessarily because there are no data at the moment and so how does how do the authors do that they do that in a creative way because they look at existing programs that are closest to the RF that is the European Regional Development Fund and the European Social Fund and so they they do the estimations and of the of the subsidy and what is the transfer and pulses on the number of of variables and given that the RF is relatively close to these funds we may you know this this says something we expect this to say something on the effects of RF they also try to analyze further by setting up a New Keynesian two two country model with you know with R&D spending and education spending so as to you know further trace out the potential effects of of the RF so I think too important well what is crucial in this paper are two important elements and to what extent does the composition of the two existing regional funds and match that of the RF and to what extent does the theoretical framework capture the main features of the RF I think these are two crucial questions on which I want to say something before doing that a few words on the empirical analysis the paper is and not only original in addressing the question so what are the effects of the RF but it also is one of the few papers to use regional data and I think the the regional data are very rich but they are underexplored and there are some exceptions a recent paper by Sebastian Hautmeier last year which looks at the regional effects of monetary policy shocks and Jacopo and another co-author we have we are working on a central fiscal capacity paper where we also use regional data now of course there may be endogenous effect from economic outcomes to the transfers that regions receive so that is for that reason the authors deploy a two-step procedure so they first regress the real structural funds on the constant and some other variables and then what they do they include the residual from this regression in you know the regression that is the focus and you know for which they show the the impulse responses and they they have something like 290 regions so they can also do a number of the sections of the data for example north versus south but also based on average income so it's a very rich exercise I think the theoretical framework was briefly explained but but well explained in one aspect I will go into that is that the taxes are collected at the local level then they are transferred to the federal level and all this spending is done at federal level in the in the model I think the paper could provide a little bit more information just basic statistical information some key macro variables for the regions the transfers received by the regions so I see the map but one of the questions and I'm not familiar with these transfer data so I you know it's I may be mistaken but I was wondering you know whether there are many observations with zero transfers and whether that has has some effect of on the empirics and I think it would be useful to say more and this is probably quite kind of well laborious you know exercise to say more on the composition of recovery and resilience plans as so what is exactly in it in those plans and these plans they differ of course from country to country but some information on this and you know how that corresponds to the to the other to the existing funds would be would be useful I think well so the effects of the transfers both in the RF but also in the funds that are studied in detail by the authors they could run via supply and the demand side and so I think these are mostly discussed in in in terms of well supply but they can also have an effect by demand side and that is also the case in the theoretical model so I was wondering whether we could how you say you know whether they could do something to disentangle these two channels and and so one possibility would be to take the basic regression that well that okay that that was a few slides ago the basic regression and interact the transfers with slack in the economy or maybe whether the economy is the effect is at the effective lower bound and so get an idea whether the transfers are more effective when there is a lot of over capacity and so the transfers have of course a supply effect via knowledge accumulation or investment but the question is also okay you know what is the how does the effectiveness of them depend on the on the situation of the business cycle and you know a number of findings struck me because I mean one of the funds that is the ERDF focuses more on spending on innovation research digital agenda and I was struck by the fact that the positive effects die out so quickly which would suggest that and these are also investments so I was quite surprised about this I was quite surprised about the difference of the effects for the two types of for the two funds and I think I would also be good to know a little bit more about what is driving the differences between the estimates for the for the entire EU and for the euro area because I saw in the tables I saw quite large differences I think especially what is the ERDF for EU versus euro well the euro area is of course about 90% of the EU economy so I was wondering how this can be explained I thought the findings on the you know when you kind of you know dissect the data by by quartile of income I thought that was very interesting and so the middle quartiles they benefit most from the from the from the impulses a few other maybe smaller comments but let me let me go to my final two slides that is there are two in the way two important mappings in the paper had to say something about what are the effects of RF I think it is important to have a bit more explicit mapping between the existing funds that are investigated and the RF so a more explicit comparison of the composition of the spandex and of course I mean maybe it would be good to look at one or a couple of RFs because these are very you know big large you know a number of pages so to say to what extent does the financing coincide and maybe quite importantly and then I'm coming back to the state of the economy probably I mean the RF these these plans they come into existence at rather specific moments in time because this is you know when we are coming out of the COVID crisis hopefully coming out of the COVID crisis while the estimates they are of course that have been shown that of course in a way averages across all the all the business cycles then the RF contains a reform component of course a difference in that way from the from the existing funds and another difference is that the recovery plans the RF plans they are not necessarily region region-based while the the funds that are studied in the paper they are based on the on the regions now finally my final issue is question is about the mapping between the theoretical framework that is presented which I like very much and the RF and of course as I was saying the RF is both about investment and reform of course you know to model reform will be very difficult in the in the theoretical framework but in a way the model is more suitable for the RF than for the funds had the ERDF and the ESF because it is a model not of of regions but of two countries and so I was wondering whether you could potentially take a more direct approach and you know go to the RF look at the components of some of the plans and then calibrate the model directly to the main components of the RF on the model I think it might be interesting to also have some differentiation between the two countries and we know that the well the financing of the RF is of course unequally unequally distributed so you could maybe assume in your theoretical model a poorer and a richer country than do some you know impulse response analysis so let me stop here and and over back to Klaus yeah thank you very much so it was very good discussion focusing on the paper many many interesting additional aspects I would we have a number of questions in the chat and I will try to summarize them and give first the floor to to Evie and then to really and I think it goes also all the questions together and then you can see where you would like to apply or focus on a first issue is more a clarification that the toilet Dorucci stresses that there is of course a difference between new generation you and this and these existing funds Evie which you have investigated namely that now there is a lot of effort to implement the proper governance structure first completion of milestones and so on so how would you see that this difference between the governance difference between the RLF and the old funds which you have investigated or the previous funds could be be addressed then Marco has say more technical question what you think about human capital depreciation which normally is assumed to be slower than because you have learning by doing and so on that depreciation of physical capital and Francesco Sonetti asked the question I think that's I have also thought about this whether you have looked at same differences in the policy the quality or administration quality of administration quality of institutions cause at different regions there is some literature which tries to link poor quality has then certain spending has less effects but when I look at at your result I would think that it's perhaps not not so simple and perhaps for you if you could turn it around your results and say okay those regions which are very successful using these funds maybe they have a better quality of administration or institutions than we normally think or that could be a check of what the World Bank or the the other say the other institutions which assess quality of governance are doing so in that in that respect there is another question by Jacopo who says whether you control in step one and two for national and regional fiscal policy and I would want to add an own observation from my perspective they have you try to say every region there is a catching up process and you take an average catching up across regions you assume that certain that every region would catch up with a certain speed to the say to the euro area every level the best performance in terms of productivity or GDP and then you you assess how these funds would would impact on on a on this speed of catching up two countries or regions so the regions which get the funds catch up faster or less fast than what you would on average assume giving that position in the productivity level or the majority of the economy that would be in the perspective which you could also bring in so sorry I now added many of the all the questions together and let's see what what you what you reply is if you well first of all I would like to thank you all for your attention and your comments and more more and hey buddy royal because like more I have here two pages of comments and it's very good so let me tell you like something on the basic ones so about the mapping of RRF and here at the FNSF I'm with you what I could do in empirically at least is that I could look at the funds and try like to see which ones really correspond exactly to what is like RRF it can it can be considered as RRF and repeat my regressions only for these funds that they could be thought only as RRF and this is something that we can do and it's easy for this then I think you gave me a very good suggestion that could be another paper which is okay take your theoretical model it looks like it's good for RRF and try to calibrate like your model and try to talk about the RRF looking at your theoretical economy I think that I mean I think that if this is a very good suggestion but I don't think it is it is going to be I think you're to do everything in this model because already in this model we're doing too much I mean this paper is like you have seen it it's like very long and I don't want to make it like longer however I have to say that the model helped me in some directions so for example you told me before that you know you don't understand why the RRF effects dissipate so fast and actually you also told me you know I don't remember if it was you somebody else I don't remember that it is there's a difference it was you between the Euro area and the EU regions so it looks like that this is driven by the port and actually the newcomers now what happens with the newcomers something that we were not aware of when we were writing the paper and we have been told afterwards for the ERDF funds the new accession countries have the possibility if they don't use the ERDF funds for their for their scope they can use them for other uses this is what they would have been told so for example in Bulgaria if you don't use the money to create like to subsidize new firms you can use them to fix payments so it looks like that this effect of the ERDF that you know if they dissipate so fast it could come from the administration of the funds now the way we saw this in the model is by looking at these reversals so what is the reversals basically what we want to do is like we say look in order for the fund to have effects it needs to be persistent you know I'm not going to invest in capital if I know that the ERDF the call for patents for the EU is here today and tomorrow is going to be gone because this is this is going to definitely generate an reversal so in the model what we say is that whenever you have this kind of reversal what is going to happen is that the effect of the shock is going to to dissipate very, very fast now I want to go now to talk a little bit of the capital depreciation because I think it's very important it is true one would think that you know human capital depreciation should be much smaller and actually we calibrated to be much smaller but I think that in a very stylized model I mean stylized whatever it's very big but anyways in a model like ours which we cannot actually model everything what human capital depreciation stands for is migration of skilled workers of human capital so in this accession countries we have a lot of migration of high of high skill agents and also like in Greece I mean I come from there I know so like there is brain drain and the way that we could capture brain drain in our model is through this depreciation of human capital and if we do so we do actually find that if this human capital depreciation is very high obviously the effect of these HK funds the human capital funds is dissipating also fast because basically you give them money somewhere and at the end like this human capital is just leaving your court okay so now institutions institutions is something which is very important and the only thing that we can do is that we will cluster our units at country level because we think at the country level the institution are similar now Francesco will tell me go to Naples and go to Milan and compare yes I agree but there is not too much I can do because I don't have any data on institutions what I could do maybe but I mean this is going to be with like recent data is to use the euro barometer information about how like questions that they have there like how happy you are like with the administration and the governors but I mean I don't have data for the 1980s for these things like all like I could use like more recent data but it's something I could try to to address this way Jacopo thank you very much for the comment yes we have we are we have some regressions in which we control for national fiscal policy not not regional and also we have also some experiments in which we also control for national income since national income bachelors EU income they're kind of correlated business cycles are correlated it doesn't make a big difference cuts up Klaus thank you very much but we haven't looked at this made I think the only way we can think about cuts up is like looking at the quantiles that we are doing currently I mean I need to think a little bit more about how we can do it in another way and finally a very important comment from Royal that I really appreciated you know Royal I also thought about it is about this non-linearities so Royal said Evie look I mean we're talking about the RRF but the RRF are here because we are under the COVID and we want to see how they are going to affect the economy given that we are in a recession so it would be great if we could do like state dependence and look at the zero load bound or look at the recession versus expansions but we only have 28 years of final data so like with the data that we have we have done I think the best we could so I will leave it here I left you three minutes there or I'll come off yeah wonderful thank you I mean fascinating that you addressed really all the questions Evie that was was very good and now we go to the Royal I mean I you have given that we we have no no discussion in the next session you have yeah two, three minutes well well okay thank you Klaus let me say a few few words I thought there was some very interesting things in the reply by Evie and that is on the issue of of migration of skilled workers so I think that it is important to think very gively about how you say the design of these of these transfers and you know when I think but when I think of the RF is I see the RF as a in a way a patchwork of national plans and have because they are designed at the national level and you know what comes out of it is in a way somewhat well somewhat arbitrary configuration of plans well actually the kind of the spill over effects are not really explicit part of the of the plans and so I thought that you know thinking about the spill over effects of funds I think it's it's very very important just on the on the regressions themselves going back I can see there are few well 30 30 years of data but still I think one could probably split this sector data set into say you know above average and below average growth and do a you know tentative regression to see whether the effects of the funds are are different maybe they are not but you know given the kind of the specific moment that the RF you know funds become available it's it might be useful to do such a thing and to say a bit more about you know kind of of the comparison between the work on the two funds and the RF thank you thank you thank you very much there's another question but I think perhaps better to maybe to to address it to discuss bilaterally with Luis from SECA on and the genetic concerns but if you if you if you have to say last say last minute if you and then we comment on everything and also the rule set and then we go to the next session yes and the genetic concerns are getting what in the egressions yeah and yeah this is from from Luis from SECA who said that the methodology cleans the the fund attribution from common EU level factors and uses of the residuals my concern is that I would think that it's exactly those residuals that are more likely to be correlated with regional unobservables making them not the best instruments wouldn't actually the variation driven by common EU factors and thus independent of individual region be the better variation to use as an instrument well I mean we have I thought that like the endogeneity would come from the fact that you know this this kind of funds come like like at in in like times where like they're you know I would trying to to control endogeneity thinking like about whether the funds were going during good times and these good times were actually capturing this positive multiplier now that's exactly what I thought that we were trying to do we were trying to to get the the residual that would clean like that would have like only the original component that it would be clean for for other stuff and and and use that not in relation to anything but in relation to the fund so we were using like this residual as an instrument for the fund that was given in the specific region so I am not sure I understand why this is going to be problematic or like that we're capturing an idiocycletic component of the of the region like is he afraid that this specific region had a specific characteristic that got this fund like is this the concern like if this is what I understand and then I would say as long as this not related to the business cycle conditions I think that we should be fine okay very good I mean I think you can continue discussing this Luis from second this this bilaterally okay by email because we have reached the end of our first session in this in this okay thank you thank you very much Evie thank you all I found a fascinating discussion and really good that that you're looking at the data and the past experience to get some idea what could happen with the LRF funds and and of course the arguments about the institution and what Ettore said about governance will play a major role