 I'm going to talk about trying to put a bunch of work that I've done together in sort of a policy framework, if you like. That's the short note I said, Moritz, for this conference. The main question that I want to discuss and think about is, I'm saying here how to think about our final, but you could equally ask the question of how to think about credit. So we have this notion of credit that we obviously feel is very important in the economy, but where does credit actually come from? And that's the question that I want to think about a little bit more carefully. And I think once you think about that question and you look at the data, you come to some conclusions that I'll talk about that are quite important in thinking about where we are heading going forward and what policy sort of implication it has. So that's the key thing I want to talk about. Just to set things up, there are two ways to think of credit. One is in the long run, sort of in steady state, if you like. That's the notion that I want to have in mind today when we think about credit. So while I will talk about the business cycle fluctuations in credit, but I'm only going to talk about that if it helps us learn something about the longer term trends, the steady state trends, if you like. So when I say where does credit come from, I'm really talking about where does it come from in the steady state notion of the word. So we're interested, I'm more interested today in talking about the long term trends and what has happened to credit in the long term. So we're going to start off with a picture that is very much motivated by the work of Jordan Shodoric and Taylor. And this is a total credit to GDP. I'm plotting you this figure for the US, but you could plot it for the advanced economies of the world. The basic picture does not change as, again, the GST team has shown us already. The one difference, perhaps, that I'm doing is that I'm talking about in the left panel, I'm showing you total credit. So that has three pieces to it. Household, non-financial firm, and government credit as well. And for reasons that hopefully will come clear, I want to combine all those three. That's like total credit in the economy. Obviously, the definition of credit is always gross. There's no such thing as net credit, which is always zero by definition. But this is what has been happening to gross credit over time. And the main fact is what GST called the hockey stick fact that since the 80s, total credit to GDP has basically skyrocketed. And the other fact that I want to highlight a little bit is that if you split the private credit portion of the left panel into household and non-financial corporate credit, essentially the big growth since the 80s is in household credit. And I'm adding government piece to it as well. Just to let you, the reason I'm doing, I'm adding household and government is for reasons I'll argue later on. This sort of debt is essentially financing the demand side of the economy. And so that's why I kind of think of these two institutes of one another in that sense. So that's what I'm adding those to. But again, the main fact is that credit has grown a lot as a share of GDP. And that's sort of a curious fact. The next thing I want to do is do exactly what Eugene asked us to do, which is not just look at quantities, but look at prices as well. So let's look at the price of credit. Credit to GDP has increased by 120 percentage points of GDP. It's a huge increase. I just want to keep the magnitudes in mind. The interesting fact is that despite that huge growth in the quantity of credit, when you look at the price of credit that has, which is the interest rate essentially, that has gone in the other direction. Now I'm showing you the 10-year real interest rate. That's sort of the most liquid market for long-term debt. And it's unquestionable and not really a contested point that globally, interest rates have gone down and gone down significantly to the extent that today, for a while, actually, we have been hovering around in real terms an interest rate close to zero. And moreover, it's been extremely difficult to get out of this. Even today, if you look at the news, that's the everyone has these expectations that these long-term interest rates are not going anywhere. So if I just told you these two facts, you already learn quite a lot from basic one-on-one kind of intuition. These are long-term trends. Again, you can think of them. These are steady-state trends. If you put the quantity and the price picture together, collectively, they are telling us that for some reason, and that's the next thing I want to get into, there has been an expansion in the supply of credit. Supply equals demand in equilibrium. We all know that. But there is this notion of what is the primitive here. So from that perspective, given the co-movement of quantity and prices, it's suggestive of an interpretation that is the supply of credit that's moving first, and it's trying to create its own demand, if you like. So what is the source of this, again, long-term interest in long-term trends? What is the source of this long-term growth, long-term increase in credit supply? One possible reason behind this might be the rising inequality around the world, not just the US, and here is one illustration of that. This is from the world inequality database by Piketty and others. This is the top 1% share across the countries that we have data on over time. And you can see that the picture that we are more familiar with in terms of the US picture is basically on average true for these advanced economies that the data set follows. In particular, the quality has measured by top 1% income share. This is pre-tax income share. It's important, given what I'm going to show you next. That came down after World War, but then again, the timing is exactly the same. From 1980s onwards, that has increased significantly and substantially. Interestingly, it's not just the pre-tax income share of the top 1%, the other thing that has happened is that the top marginal tax rate for exactly that bracket, that cohort of people, that has declined a lot around the same period. So again, now I'm showing you the top marginal income tax rate. The two lines, by the way, one is the simple average of these countries, the other one is a weighted average by population. The basic trends are the same. You can see that the top marginal income tax rate has declined on average quite a lot. This is a decline of 20 percentage points or so. That's a very significant change, and exactly it happens around 1980s. As has been documented in the inequality literature, when we talk about rising inequality, by and large, it's a top 1% phenomenon. So that's why I'm focusing on this particular cohort of people. The question is how is this related to the growth in credit, which again, if you look at quantities and prices, suggests that it's driven by an expansion in credit supply. Well, the basic fact about saving behavior that has been extremely well documented is that when you look at saving behavior, preferences look non-homothetic, which is to say that people at very high income levels, they save substantially more as a fraction of the income that they are getting. So if you just, even in a simple sort of spender-saver model, if you were to incorporate that key fact that top 1% are very high savings rates, what you're going to have in the economy is that the pool, the gross pool of surplus savings, if you like, that is available to the economy to be lent somewhere is going to grow, right? And that's my conceptualization of the financial sector since 1980, that more and more funds are available, which are channeled back into the financial sector to go somewhere, right? And the basic sense, this supply is coming from this, in this story at least, I haven't shown you that there is a connection necessarily between the two, but one possibility is the time series is consistent with each other, that is the source of funding if you like, that is then trying to find a way in the economy. Now, at the conceptual level, I want to highlight one point, which I think is important in terms of the typical macro model that we sort of use in thinking about steady-state models. When we talk about credit in a typical traditional macro model, the notion of credit is that it's driven by some notion of the permanent income hypothesis, which is to say that credit is never a primitive force in those models. People, for example, have lifecycle motives, so some of them who are going to earn more later on, they're going to demand credit today, and others are going to be providing their savings to them for borrowing, and that basically closes the model. So there is no notion of what I just said, which is that there might be surplus savings available in the economy. To generate that notion, you need to deviate from the traditional models by introducing something like non-homothetic preferences, where shifts in income distribution will become a primitive that changes the availability of credit in the economy. But when you look at micro data and you actually look at the savings areas of individuals and you combine that with the movements and inequality, the macro model that seems more realistic is more in line with what I'm just describing here. This notion that the growth in credit is connected with rising inequality is also consistent with who has been borrowing more than others. So here, I am plotting the, and this is using public tax record data. So this says Zuckman tax record data where they impute household debt using tax deductible interest payments. And if you plot the leverage ratio of households, again, I'm going to focus on probably 1% versus the rest. So that's where inequality is, or change in inequality is. What you see is that this, again, we are familiar with in the aggregate what is happening. The hockey stake, if you like, that starts around 1980. That big increase is essentially all of it is being driven by the bottom 99%. It's not there in the top 90, top 1%. One can open up this piece further and you generally see the same picture that the credit is coming from where more and more income is going and flowing in the rest of the population. So the cross-section of debt to income is consistent, again, with this motion that credit is supplies flowing from the top of the income distribution to the rest of the economy. Now, there is another variable that is also acts like an expansion in credit supply to the economy, and that is this motion of global savings blood. So everything that I've talked about so far would hold true in a closed economy. In a closed economy, rise in income inequality, all else equal has to show up as credit somewhere else. But in an open economy, there is another degree of freedom, which is you could export out your excess savings to someone else in the economy. Number one, number two, in terms of credit absorption, there is this substitutability which I highlighted earlier, which is that if there is more credit available, either households could borrow it. But if for some reasons, household cannot or do not borrow it, the government sector can step in and they can borrow on behalf of the households. It was for that reason that I was combining household and government credit together in the total credit graphs that I was showing you earlier. So on this graph, what I'm showing you is for each country, over this time period from 1980s to basically present, I am integrating the current account or net exports if you like, plus the deficits run by governments. So this is the net absorption of credit if you like in a country during this time period. And on the other axis, I'm plotting the change in stock of household credit. And what is interesting is that you see a very strong negative correlation between the two. So what you can see here is that countries, so one way to interpret this is that countries, it's like the global savings club. So if you focus on the net exports here, countries that have been running current account deficits, they have seen their household debt run, grow faster, all else equal. Another way of understanding the same is that countries that have been running large government deficits, they have the government absorbing the excess credit that is generated. And so household debt does not need to grow as fast as it otherwise may have. So all of these variables are linked together if you write down a basic macro model, which is how we are deriving these empirical relationships. So we can put all of these facts together at the macro level in an open economy macro sense. And now we can ask the question of, does movements in inequality times the top marginal interest rate, not interest rate, top marginal tax rate changes from 1980s to 2000s? Do they explain the rise in household credit or not? Controlling for these other two key factors, is this clear? So that's what I'm going to show you next. Controlling for these factors, do changes in inequality. So I'm going to essentially assume that changes in inequality are being driven exogenously for some reason. I don't have an instrument, and certainly I don't have the number of observations to do anything more sophisticated. I'm just going to show you a simple plot using this cross-country data. So that is what is shown here, where I'm basically partialing out the other variables of interest that come out of the macro model. And now I'm showing you the relevant partial correlation between rise in inequality and rise in household debt. And what you see is that basically there is a reasonably strong correlation between the two in this sample of around, this is at least 17 countries that we have consistent data on over this time period. The broader point I'm trying to make here is that there is this connection in the data, both in the cross country level, as well as at the micro level, if you look at the cross section of who's borrowing within a given country, there as well as the time series, which is what I showed you initially, this connection between rising inequality and rising levels of household credit. So all of this suggests that this is one of the fundamental structural shifts since 1980 that a lot of people have talked about, growth being unequal. That's a broader lesson to take away from all of the discussion that we just had. That since 1980s, the global growth process has been unequal, among other things, given rise to this top 1% share. And that has led to this greater aggregate availability of gross savings that shows up as credit to the financial sector. OK, now the next question is what are the macro ramifications of a pattern of this sort, a trend of this sort? Let's go back to the basic traditional macro models. In basic traditional macro models, all else equal if I told you there's more gross savings available, you'll say that's great. That means more investment. That means much higher growth. But that's not what has happened. And so I think that's the other dimension that we need to deviate from the traditional model. So it was actually in the macro framework that I was talking about earlier, you could have asked me this question of when I was showing where could this trend be absorbed, could it be absorbed by government deficit, you could have asked the question, well, it can also be absorbed by the capital formation sector. It could be absorbed as real investments. And so you could account for that empirically, and we do. But the basic empirical fact is investment to GDP has not increased at all from 1980s onwards. That's just so obvious that I don't even have to prove it. Given investment to GDP has not increased, in fact, some argue that it has gone down somewhat. Given investment to GDP that's real investment to GDP has not increased over time, credit to GDP has increased a lot. Again, the question comes back, where has it all gone? What is being financed by credit? And that's the central question. What I'm going to argue is that when you look at the data that suggests that credit increasingly since the 1980s, unlike our traditional textbook models, it's not the increase the marginal credit, if you like, it's not being used to finance investments, real investments, but it's being used to finance consumption, what you might call credit as demand. So this is what in a recent jet piece we call the credit-driven household demand channel. We are limited in terms of proving whether this kind of a channel exists by just looking at very long-term transfers because of the usual problems of identification and so on. However, I think the business cycle facts are interesting in convincing or useful in convincing us of this particular channel. What we have just talked about in a way is the super cycle that is in the background to a number of the business cycles that we have seen since 1980. So in other work with Emil Werner, we show that if you look at business cycles around the world since 1980s, they share many of the features that we saw in 2008, the crisis in 2008. And I'm summarizing some of that information here. And using that data, we can sort of come to this view that it seems more and more recently that credit is being driven, credit is being used to drive aggregate demand. And that's sort of an important feature of the macroeconomy. So let's look at or let's summarize that evidence that leads to this conclusion of credit-driven household demand channel. On the left panel, I'm summarizing what I think is one important fact from the Great Recession, which is each dot is a state in the US. When you sort states by growth in household credit GDP during the boom years, or two to seven, that predicts, in a statistical sense, the output gap, which is the rise in unemployment in the same state from 2007 to 2010. Interestingly, if you were to plot exactly the same graph but cross-country rather than within states, you see exactly the same pattern. Even in terms of empirical magnitude, the coefficient is comparable. What we argue in this paper, and I'm going to show you some evidence using a natural experiment from the 1980s, that this pattern is not only true conditioning on the recession. I'm cheating here, obviously, because I'm conditioning the scatterplots on the recession. But even unconditionally, this relationship holds true, which, again, means that stronger growth in household credit unconditionally predicts slowdown in GDP growth going forward. The question is, what is moving credit around? I've already talked about movements like inequality is providing an expansion in credit supply in the economy. Disruptions in that process and the business cycle frequency then move aggregate demand up or down. And that's what causes business cycle movements. What is the proof for that? Well, the proof for that is that when you isolate these business cycle movements in credit supply, that I'm going to talk about one natural experiment that does that, what you see is that that movement in credit supply moves aggregate demand up or down. And it is, when it's moving up, it moves aggregate demand up. And on the flip side, it is that expansion in aggregate demand that then predicts future decline in GDP because of the success supply of credit. So what's the proof that expansion in credit supply and the business cycle frequency is moving aggregate demand? Well, for that, let me just use the 1980s as a natural experiment. It's something that all of us who have read Phil's strands work are very well familiar with deregulation of bank branching network at the state level or staggered. So we can use the timing of deregulation as a natural experiment at the state level. That's I'm just using the strand experiment to run that natural experiment. This is the first stage, early deregulating states saw a larger credit boom than late deregulating states. So the gap between the blue and the red is the first stage. The pre-trends are similar. That's basically the exclusion restriction. Again, that's not our work, has been well documented before. But what I'm going to focus on is that first of all, that credit boom has something very similar predictability that I showed you for 2008 as well, which is that it leads to a boom during the credit boom years but then a bust afterwards. So during the credit boom years, states that see a larger boom in credit, they also see a larger boom in real outcomes. So a larger decline in unemployment rate if you like. More interestingly, for our purposes, we want to understand what's driving this boom. And the claim I want to make is, this is the key question given the longer time series that I showed you, what is incremental credit at the macroeconomic level? What is the incremental credit being used for? And I'm going to argue here that that incremental credit is being used to move aggregate demand up and down. So how can we separate whether it's moving aggregate demand or aggregate supply? Karan, I changed the paper completely, so. So how can we test that it's moving aggregate demand and not aggregate supply? The key notion again comes from one-on-one, which is let's look at quantities and prices. And in particular, look at sectors that move differentially in terms of demand side versus supply side. So we're going to focus on the non-tradable versus tradable sectors. The basic intuition being that if a given variable here, it happens to be credit, is primarily moving the aggregate demand side, then it will disproportionately impact the non-tradable sector in that local economy. Because it's the non-tradable sector that, by definition, depends on local demand disproportionately. So relative to the tradable sector that caters to international or national demand. So if this credit variable is really moving local demand more than local supply, then it should primarily load up on the non-tradable sector, both in terms of quantity as well as in terms of prices. We should see an increase in relative quantity as well as relative prices of the non-tradable sector. So that's what I'm going to show you next. The x-axis, I'm sorting the same states by the extent of the shock, which is the timing of deregulation if you like. So that's the instrument. And on the y-axis, I am first showing you the aggregate, sorry, this is the predictability graph. So this is the same graph that I showed you in the very beginning. For 2008, states that levered up more suffered a deeper recession. You see the same pattern in the 1990s recession. So first, I just want to highlight the similarity. What we saw in 2008 in the US as well as across country to see a very similar picture in the 1990 recession. It's just that we didn't talk about it as much because it wasn't as salient as a subprime crisis. But let's focus on the boom, which is what I'm trying to use to discriminate whether it's moving aggregate demand versus aggregate supply. I already gave you the hypothesis. Let's look at non-tradable versus tradable sector. Again, the x-axis is the same, sorted by the instrument. What you see is that it's the non-tradable sector that expands disproportionately more in states that are experiencing a larger credit boom. The tradable sector, you don't see that at all. And also importantly, we can focus on non-tradable prices as well. And it's the non-tradable prices that are appreciating more. So you're seeing great exchange rate appreciation, if you like. Because these are states in a monetary union. Tradable good sector prices are not differentially moving. So putting this non-tradable sector evidence together, it suggests that credit was moving aggregate demand around. All right, let me close now to think of what does all of this mean. So let's now move forward. And let's say we believe in this macro model where this global savings blood rising inequality is moving credit supply higher and higher. And prices, which is the interest rate, is now adjusting to accommodate more and more credit in the economy. Where does all of this end? Everything I've said so far is not a problem for a macroeconomist, because macroeconomists don't care about distributions. So as a macroeconomist, everything I've said is just fine. Prices are adjusting, which is like the interest rate is adjusting. And demand is now being funded differently. But who cares? The aggregate economy is still the same, except that there is a problem. And that's what I want to focus on in the end. There is a problem if you start running against certain aggregate constraints. And there are two types of constraints that I want to talk about. I'm going to just talk about those two constraints through the lens of a simple graph. So let's write down the usual supply and demand curve, but in the growth interest rate space. So in the growth interest rate space, the demand curve would be our traditional Euler equation. And the supply curve could be whatever we want to make it to be. So in the traditional solar model, it's exogenously given that it would be a horizontal line. A more endogenous model would say that decline in interest rate makes firms invest more, so that raises interest rates. That would be a downward sloping line like that. The key issue is that if given everything that I've talked about, if you think that it's rising inequality that's driving movements in interest rates, the first set of graph that I talked about, then it's the Euler equation that's shifting to the left. And the question is, what does that do to the aggregate economy? Again, in steady state, let's forget about transition dynamics. If we are in a solar model, and it's just a horizontal line throughout, as I said, there's nothing to worry about. Interest rate are just, and we are where we would have been anyways in terms of the growth rate of the H1. The question is, what if that is not the case? And there are two reasons to worry about that. That might not be the sort of the solar model intuition might not hold at some point, at some margin. The first one has been talked about a lot, which is this problem of zero lower bound if you like. Another way to talk about that is that if it takes ever lower interest rates to convince people to borrow more and more, what if you hit zero, and now there is no more price adjustment to be had, then you hit what is called a liquidity trap. You cannot convince more people to borrow, and that results in a liquidity trap. In the liquidity trap world, the only way to absorb the surplus savings is for the economy, the real economy, to contract to shrink. That's exactly the definition of liquidity trap. So you could think of the economy now being close to that liquidity trap point, and growth must suffer to accommodate these increasing surplus as the cause of liquidity trap. That's one possibility. The other possibility is that even before we hit liquidity trap literally at zero, that at very low interest rates, there are other problems that arise from very low interest rates. And one possibility that we talked about in the paper with Ernest Liu is that at very low interest rates, markets become less competitive or more monopolistic. And as a result of that, growth actually reduces. So you have this backward bending supply relationship, and as a result of that, as interest rate continues to get close to zero, growth actually slows down. The broader point is that if that is the world that we are increasingly living in, then we need a more equitable growth to get rid of these surplus savings. B, this is coming from an earlier work that you need to worry about risk-sharing across these spenders-savers in the economy. And finally, the most radical conclusion that comes out of this sort of analysis is that wealth tax might be useful, because again, you're trying to absorb surplus savings away from the economy, and then at that. Tief was right. He did change the paper that he was discussing. You have the paper in your packet, and a lot of the themes that he was just talking about are just really related to what's in this paper. So I actually do have a summary of the paper, at least the way I read it at the beginning, so this will help orient you. In the paper, Tief talks about this enormous increase we've seen in credit relative to GDP in recent decades, and he has a chart showing that the increase is mainly explained by a rise in household and government credit. He talks about how a central cause of this increase in credit has been a build-up in what he calls financial surpluses, but basically credit supply, which has come hand-in-hand with a decline in the equilibrium interest rate. He notes that higher income and quality has contributed to this build-up, with the logic there being, of course, that rich people have higher propensities to save, so if you shift income towards them, you get more aggregate saving. He talks about how these developments have various bad implications, they'll come back on that in point, and he talks about ideas that can help, including more state-contingent debt contracts, which would be the risk-sharing products that he talked about at the end of his talk. Just as my meta summary, I wanna say, this paper's a great paper. You should definitely read it. It's very thought-provoking, so that's my meta comment. I'm gonna give some specific comments in several areas. I'm gonna talk about the importance of the topic. I'm gonna talk about this idea of productive and unproductive credit. I'm gonna talk about how well the story in this paper fits the current situation, and then I'll get into some policy issues about better protecting household borrowers and other policies that can help. Okay, so I'm gonna start with the importance of the topic. I think this topic is hugely important, so I'm just grateful that he wrote this paper. I think it's among the most pressing issues that we're facing in macroeconomics today. It's strongly linked to the secular stagnation literature. I should say I was kind of a lapsed macroeconomist for a while, actually for a number of years, because I just wasn't really compelled to be thinking about it during the period of the Great Moderation. It just didn't seem that interesting, but I'm back to kind of thinking and researching in the area, because I do think these issues are so important right now. But anyway, kind of like the key thrust of the paper is as they would say in Econ 101, there's been this great growth in the supply of loanable funds relative to the demand for loanable funds, and that's caused this big increase in the decline, or a big, sorry, a big decline in the equilibrium price for loanable funds, which we would call the neutral real interest rate. And just to give you a sense of the magnitudes, this is a screenshot from a paper by Larry Summers and Lucas Rochelle that was at Brookings last spring, but it's a screenshot showing you the decline in real interest rates in the US and globally since 1980, and you can see it's on the order of three to four percentage points, and that is what, indeed, the kind of literature is estimating as for the decline in the neutral rate. Anyway, these developments have led to a set of serious concerns. A key theme of this paper, and it's what Artif was talking about at the end of this last presentation, is that we perhaps live in this world where we are unable to get to full employment without all this debt, which is exposing the borrowers, and it's exposing our economy to risks that make it unsustainable. Yeah, there are other bad things about it as well. Many people think that low interest rates can make bubbles and broader financial instability, more likely, low interest rates, particularly if they're unexpected, create problems for insurance and pension providers who are on the hook for payments down the road, and then as Artif was saying at the kind of, end of his presentation, this is a really big one, low interest rates limit counter-cyclical monetary policy, so this is a zero lower bound problem, and it is, as a macro economist, I find this probably the most scariest implication of this all, the Fed came out yesterday and had cut their neutral policy rate again to 2.5%, you consider the fact that in the last three recessions the Fed has cut rates by five percentage points to stabilize the economy, we're in big trouble next time around. Anyway, let me move on to my next comment, which is on productive and unproductive credit, and I have to say, one of the things I like in this paper, and I think it's in the paper you just presented, is this idea that there's credit that finances things that expand the supply side of the economy, and then there's credit that expands the demand side of the economy. I do think in this particular, at least in this paper, they talk about how kind of credit on the business side is the productive credit, because it increases the supply side of the economy, I want to kind of push that back on that a little bit as an oversimplification. I do think there are other types of credit that can be productive for the economy, so even though I do a lot of speaking about how there are problems with the US federal government student loan program, and there are problems that need to be addressed, they are used to make important investment in human capital, and that will increase the supply side of the economy, and there are papers, for example, worked by Michael Greenstone and Adam Looney showing that actually for most students investing in a higher education, even with student loans, is going to give them a bigger return than other types of investments. Mortgage credit, we know there are definitely risks associated, but I will kind of point out that traditional mortgages, they're basically this commitment device that can help households overcome behavioral barriers to saving. So that can expand the supply side of the economy by giving households the saving that they need to store businesses or to finance higher education. And then, of course, government spending, that can be viewed as an investment, or some of it can be viewed as an investment as well. I'm not just talking about infrastructure and research, which we know should explain the supply side, but there's this great new literature which is showing that kind of spending on social programs, programs that are aimed at making the lives of children better have a long-term payoff in terms of making these children more productive as adults. So that's another way in which government spending can increase the supply side of the economy. As long as we're on the issue of kind of different types of credit, actually one thing that's not discussed in this paper, but I'd love to get thoughts on is this big increase we've seen in financial credit. The graphs that Ateef showed were actually just domestic non-financial credit. That's kind of one of the headline numbers in the flow of funds accounts, but as I was poking around in the data myself, I was looking at the ratio of financial credit to GDP, and I was noticing how Ateef's up 60 basis points over this period, so that's about two thirds as large as the rise of household and government credit combined, and I don't have a view on whether it's productive or unproductive, but I would love someone who's an expert in finance to kind of weigh in on that issue. Okay, so third comment I want to make is how well kind of this general narrative fits the current situation. Okay, so for your consideration, we're at full employment now, okay? Households are not, right now, they don't look like they're overextended. Okay, so we can see this in the aggregate debt to income ratio and in the debt service to income ratio, debt to incomes down to its 2001 level, debt service to income is kind of at the low end of its range for the last 30 years. You can look at other measures for households and none of it suggests there's a lot of trouble for households right now, okay? If anything, there are worries right now about some types of business credit. So for example, the Fed's recent financial stability report, they talk about the rapid growth in leveraged loans as a kind of area of risk. I mean, they note as Ateef does in his paper that overall businesses have not been borrowing a lot, but there are areas that we're worried about, okay? So what is going on? You have to look at the kind of key charts to figure it out, okay? I've just redrawn Ateef's charts. He probably showed them at the beginning of his presentation, you did, okay? So I love these charts, okay? So this is the chart, this is just, I just redrew what he did. It's just credit relative to GDP and then you can see over in this chart, you can see the household and government category lumped together. You can see that this is what explains the big rise in credit to GDP, kind of non-financial business credit has kind of grown moderately over the decades, okay? So what I wanna do is I wanna break out household and government credit. I've done it on this graph. I've kept the scales the same as the last graph so you can't see what's going on. So what I'm doing in this graph is I'm kind of zooming into just the recent period, okay? So this is basically the point I wanna make which is you can see the household credit as I just told you, it's come down since the financial crisis. Households don't look like they're overextended right now. What's high now is government credit relative to GDP. That's the driver of high credit relative to GDP, okay? So this is then begging the question is how do we feel about kind of this high level of government credit relative to GDP? So traditional view, a bunch of my friends back in Washington probably still hold this view is that of course it's a problem to have high government credit relative to GDP but I wanna point out there's new thinking and research on this issue that is kind of challenging that view. So this paper I already mentioned by Larry Summers and Lucas Rochelle, they kind of basically are looking at the rise in or the expansion of governments and the associated increase in government debt around the world over the last several decades. And they calculate that the real neutral interest rates would have declined by an additional four percentage points if not for the increase in these programs, okay? So if you think kind of there's been this and we know there's been this development regarding loanable funds and kind of if you're worried about how far it's driven down the neutral rate, the common would be worse if we hadn't seen this rise in government credit, okay? Other literature as well, Olivier Blanchard's presidential address, for example, talked about how with R so low, all else equal we shouldn't be so worried about this level of government debt because we've got R less than G and we'll outgrow it. So I'm not saying things are to all else equal because we've got these kind of big rises and spending on the horizon related to support for older households that we need to fix. But I am saying that the kind of current level of debt relative to GDP is maybe not so bad. Okay, so now let me get onto a couple of thoughts about policy. So first of all, better protecting household borrowers. I do think, well, not that pessimistic about the current situation. I do think there are real risks around household use of credit and around household financial fragility more broadly. So in a team's paper, he talks a bunch about state contingent mortgage contracts. I think they sound like a good idea, kind of more risk sharing between the lenders and the borrowers. Having these things would mitigate business cycle fluctuations. It would reduce the need for these ad hoc policies to rescue homeowners in the face of another mortgage crisis I gave several years of my life to that effort. Great, we don't have to go through that again. I think there are some complications around these alternative products. So paper does allude to this. Lenders do need to be compensated if they're taking on more risk. So I'd love to see more thinking about how much do they need to be compensated. And how would we feel like, because that's gonna show up as a rise in interest rates. How would we feel about that rise in interest rates? And if we don't like it, do we want a government subsidy? It was a kind of real problem when I was in government as we were trying to do GSE reform that we had kind of a deal that we thought would be a good deal. But the opponents came forward with this research that was showing that there'd be this enormous rise in interest rates for borrowers, which I don't think the research was right, but it was enough politically to kind of send the reform deal, kind of just basically kill it. So I do think it's like important when we talk about these alternative products to kind of try to put a number around how expensive they're gonna be and how are we gonna deal with that expense. Few complimentary ideas for protecting household borrowers. So this is again, along the theme of like, basically we wanna prevent them from going underwater in their first place. Okay, so I think there are mortgage products that people have suggested that amortize faster. Ed Pinto and Steve Olinner at AEI have this wealth-building home loan, which I think is really interesting and could be explored further. I think if we're interested in avoiding households kind of going underwater, we need to fight notwithstanding these zero lower bound problems, we need to fight recessions vigorously so we don't have housing markets collapse and prices collapse or over-correct and send people underwater. I think we need to preserve, of course, strong consumer financial protection. Okay, last thing is I wanna just run this out, talk about other policies that can help. The paper and a tip at the end of his talk talked about reducing income inequality to reduce credit supply to fix the situation. I agree with this point, particularly given the other downsides of income inequality. But what you wanna keep in mind is that the higher income inequality probably, according to the academic literature, only explains like 20 to 30% of the decline in the neutral real interest rate, okay? So it's less clear what, if anything, we can do about these other factors that we think are contributing. So this is kind of the demographic forces, the kind of precautionary saving, the buildup of reserves by emerging market economies, certain types of technological change that are thought to be contributing to the problem. So just kind of what else can we do? A few other priorities. So one, strong macro-predential policy. The chief talks about kind of these better mortgage contracts as a type of macro-predential policy. They agree, but there are other really important things that we should be doing. Top of my list would be that regulators need to be willing to deploy counter-cyclical capital buffers, credit booms, to stem bubbles, keep hassles out of the mortgages that they can't match. So I think that that's important with neutral low interest rates, kind of mostly here to say, notwithstanding what we do about income inequality, I think we do need more thinking to address the zero lower bound constraint on traditional monetary policy. So we need to be more willing to kind of deploy QE, we need to explore these other ideas like higher inflation targets, think about negatives. Okay, so that was basically my comments on this paper. I really enjoyed reading it. This is just my summary, you should read it too. It really is very thought-provoking. My personal view about the current situation is not as dark as the view in this paper, but I do think that there is plenty to worry about around these topics. Yeah, there's plenty on the list of kind of policies that we should be thinking about to address them. Mark? Yeah, Mark Carlson. So my impression is that sort of along with this expansion of household credit, part of this has gone to people who previously did not have access to credit. And so I was just curious, you know, how that fits in with your story and would you be concerned about sort of people being, those people losing access to credit markets if there was a dramatic cutback in the level of savings? Yes, I was thinking in terms of the cross-country evidence, just looking at household debt, so a huge determinant of the cross-country variation in household debt is how much households borrow for mortgages. And, you know, and you can see that variations in home ownership do not translate in differences across countries in households, Tennessee to accumulate wealth. But, you know, institutional factors in the mortgage market matter a lot for accumulation of household debt. So with respect to that, I think we can actually, you know, learn a lot from those institutional differences, both in terms of how to better structure mortgage products to ensure more resharing between borrowers and lenders, but also when we're comparing cross-country data, you know, those differences are going to matter for how we think about, sort of, you know, household debt to GDP ratio in one country versus another. And so probably you addressed this in the paper, but I would like to get your thoughts on this. Thank you very much, and thank you, Karen, for doing a great job, and sorry for all the trouble I caused you. Let me, I want to make two sort of points. I think partly to clarify one issue. My motivation for raising this conversation is very macro in nature, and what that means is that some of the issues that we might care about at the micro level or at the business cycle frequency are actually not that relevant. So let me push back on something that I myself said, which is this risk-sharing and state contingent contracts. It just happens that in this framework, that stuff is important for dealing with business cycle stuff. But in general, state contingent contracts or any kind of macro prudential is not going to solve the bigger problem. And I think that's very important to understand. The piece where I deviated and Karen didn't hear that piece in the beginning was that I said, okay, let's just talk about steady state. So if you're just thinking about comparing steady states from the 80s versus the more recent world, and we are worried about stuff like liquidity traps and so on, macro prudential does nothing to solve the problem. Think of it this way. If we force credit to contract through regulatory measures, economy will just slow down and it will contract. That's literally the way a liquidity trap economy would respond if we just forced less credit. Because actually liquidity trap economy actually wants more credit, not less. That's the irony, if you like, that while it causes all of these credit issues, credit itself is a symptom, not a cause of the problems. The fundamental cause, and that's partly what I was trying to get at, the fundamental cause in these kind of problems is this unequal growth that is driving the excess supply of credit in the long run in this long run steady state sense. I think that's something very important to keep in mind. And sometimes we are just, we might be sort of running after curing the symptoms, so to speak, without the underlying disease. So that's point number one. Related to that, and this is now in response to Mark's question, what I just said is also related to your particular question. It's less, again, from a macro perspective, of course it's a very important micro question, so I don't want to dilute the importance of those issues, including the one you raised, but from a macro perspective, it matters less in terms of who exactly is borrowing and so on. Let me give you one example of that. It could be that somebody is coming in and borrowing for the first time to buy a house for the first time, but in that process, and this is actually how it works, and again, if you actually put this down in a macro model, in that process, the price of housing has increased, and so they pay a higher price because lower interest rates have increased the price of housing and they pay for the higher price with credit, otherwise they would have just paid all cash, but who are they making that payment to? They are making that payment to an older generation that has now gotten a lot more cash than they otherwise would have by selling their house to this younger generation, these first time homeowners. That extra cash that gets generated will now be thrown back in the economy as additional demand. That's the key thing that matters from a macro perspective, and so when I say credit as demand, it is in that macro sense that credit generates aggregate demand. It doesn't have to be literally the same person who is buying the house, because they are buying, they are living that, and they may not literally spend that money in terms of consumption, but they are releasing liquidity for someone else who is then spending that money in the economy. So what I'm trying to say is that we need to focus on those aggregate linkages and think of what this is doing to the aggregate macro economy to fully understand the process of how credit links to the aggregate economy. Just as a side note, once you think through that, one factor that actually becomes very important is intergenerational mobility. You can have a high level of inequality of the time that we have been having, and that would be fine if you had somehow, which is not the case, very high intergenerational mobility, because a very high intergenerational mobility essentially acts like a very high top marginal tax rate, because your children are going to be very poor, and if you're very rich, and they're going to spend most of what you say, and that really helps the aggregate economy. So I just want to highlight that it's those kind of linkages that become really important from a macro perspective when you think about these courses.