Luke Bartholomew 00:00

Hello and welcome to Macro Bytes, the economics and politics podcast from abrdn. My name is Luke Bartholomew and today we are talking about the long-term drivers of interest rates. And this is a discussion that gets at the heart of a number of extremely important debates in markets and the economy at the moment, including just how tight monetary policy is, how much downward pressure it's exerting on the economy, but also where interest rates might settle once we're past this current period of concerns about recession and inflation. So, I'm delighted to say I am joined by two fantastic guests to discuss this topic. First, our very own Bob Gilhooly, who is a Senior Emerging Markets Economist at abrdn, and has recently done some research on the drivers of long-term interest rates. So, Bob, welcome to the podcast.

 

Bob Gilhooly 00:52

Hi, everyone.

 

Luke Bartholomew 00:54

And also by Lukasz Rachel, who is the Assistant Professor of Economics at UCL, University College London. He has previously had visiting positions at Princeton and Harvard Universities and before his time in academia he had a distinguished career at the Bank of England. He has produced extensive research in the past around drivers of long-term growth, long-term interest rates and the optimal policy regime with various co-authors, including Larry Summers. So, Lukasz thanks so much for joining us today.

 

Lukasz Rachel 01:29

Thanks. Great to be with you.

 

Luke Bartholomew 01:31

Okay. So, when economists talk about the drivers of long-term interest rates, they often have in mind this concept typically referred to as r*r* or equilibrium rate of interest, or the neutral rate, or the natural rate. And I think it might be possible to drive some distinctions between those concepts if you really want to but, generally speaking, they are used interchangeably and that's certainly the spirit that we'll be using them today. And the idea of r* is it's the rate of interest over the long run which will balance  desired savings with desired investment or, put in a slightly different way, it's the level of interest rates which would keep inflation stable over the long run. So, it is a concept that's deeply bound up with ideas of balance and stability in the economy. It's a pretty slow-moving idea, but it does shift over time due to anything that shifts desired savings or investments and it's important to distinguish it between policy interest rates, which of course can move quite a lot indeed, very rapidly, as we've experienced recently due to changes in the growth and inflation environment whereas r* is much slower. And indeed, we can only really tell the stance of policy, how tight or accommodative monetary policy is by comparing policy interest rates to this equilibrium concept. So, policy is tight only insofar as interest rates are above r*, and it's accommodative only insofar as interest rates are below r*.

 

And there has been a lot of discussion recently about how things that occurred during the pandemic may have fundamentally changed drivers of long-term interest rates meaning that r* is going to be higher than it was previously. And that has important implications for where market interest rates and other financial market variables will settle over time. But before we get to those things, I think it might be worth taking a step back and setting the scene on the developments and drivers of r* before the pandemic and Lukasz, you did some seminal work back in 2015 on the drivers of global interest rates and the key themes. So, could you talk us through that work, please?

 

Lukasz Rachel 03:59

Yes, absolutely. The key motivating fact that motivated a lot of the work that I and many others have done pre-pandemic on this issue is that the real interest rates that we observe in a market, the long-term real interest rates, have trended down continuously for the past three or four decades. Understanding this trend is really important from both a positive and from a policy perspective. So, we want to think about what does this trend tell us about the underlying forces in the economy, and also what the policy implications are? And in the work that I and many other macroeconomists have done in the run up to the pandemic, we kind of viewed this long-term real interest rate as essentially a price that equilibrates the asset market in the economy. What it really means is that there are some interest rates at which the desire to save in the economy is sort of met, or balanced, with the desire to invest. And so, we can think about this longer-term asset market as being equilibrated by this neutral rate of interest. And through this lens, we can then think about the changes to the equilibrium, the natural real interest rate, as resulting from forces that potentially shift both the desire to save as well as the desire to invest. And indeed, in one of the first studies that I have done on this topic, we've looked at this empirical regularity that while the interest rates that we observe in the market have trended down, the investment or saving ratios to GDP have not really moved much in advanced economies as a whole. And so one conclusion that comes out of thinking about this in this setting is that perhaps both the desire to save and the desire to invest have shifted in such a way as to leave the equilibrium quantities relatively unchanged and, at the same time, causing the real interest rate to slowly decline over time.

 

So, what might be these drivers of desire to save and desire to invest? Well, one primary driver is the growth rate of the economy. So, when people expect their incomes to be high in the future, when people expect high growth, they tend to save little today because the times ahead are very rosy. And vice versa, when people expect slower growth, they tend to save more for those times that are not so good ahead. Similarly, investment is very much dependent on the expected pace of growth. And in fact, in advanced economies as a whole, these growth rates, when we take sort of moving averages or some trend measures of these growth rates, they have declined quite substantially especially in the 

1980s. So, the 60s and 70s had really strong growth of productivity in advanced economies. And then we saw this big decline and one relatively uncontroversial reason for why interest rates have declined is this decline in productivity growth. And recently, productivity growth has also been quite weak, keeping the neutral real rates low. The other reason is the demographic change that occurs in our economies. As populations age, people tend to save more for the retirement because they expect to be in retirement for longer. And together with that comes also a fall in the population numbers. So, desire to invest is lower, because there's going to be lower workforce and therefore less capital will be needed to satisfy the production needs of the economy.

 

So, the demographics was another reason, quantitatively judged to be on par with the declining productivity growth that has contributed to this trend. Then finally, let me just wrap up with couple of other trends - rising inequality has often been discussed in this context, because the evidence shows that wealthy and rich people tend to save more than the poor. So more unequal wealth and income distribution leads to higher savings in the aggregate. There's also been a lot of work done documenting the increasing spread between the safe interest rate, the one that we're discussing today, and the return from productive assets. And that increasing spreads could have been the function of increasing financial frictions, or difficulties in financial intermediation sector post global financial crisis and that in and of itself could have pushed the safe real interest rates down.

 

I should also flag that there are some other trends that contributed to interest rates going up, at least according to the models and the theory that we have. So, increasing government debt to GDP ratios, or increasing social security spending over time, have contributed to higher real interest rates simply because some of these increases work basically to offset some of these, for example, demographic factors that are happening on the private sector side. And this balance between public sector and private sector forces is something that we have highlighted in the paper with Larry Summers that that you mentioned in the intro.

 

Luke Bartholomew 09:40

So, Lukasz when you mentioned those motivating facts for the work, it was talking of falling global interest rates since about the1980s. And what's quite striking, of course, is that 1980s is also a period of quite significant financial globalisation where capital markets became much more integrated. So, I'm wondering why r* is best thought of as a global concept and to try and make a distinction, which is hopefully not too fine. I'm wondering if the way to think of this is that there are these drivers that play out at a global level and then that just gets through capital market integration felt in every market. Or it's more that there are lots of individual domestic markets that just also happen to be driven by a common set of factors at the same time. So, it's co-movement, rather than necessarily co-determination?

 

Lukasz Rachel 10:39

Yeah, that's a great question. And I think, essentially, the short answer is it's a bit of both. So, there are certainly some common drivers across advanced economies that are playing out similarly in many countries, productivity growth has declined in several economies. There's, of course, some differences within this dimension. In the United States, for example, productivity growth has been stronger in the late 90s and early 2000s. But ultimately, the broad trend is quite similar. Demographic factors are also relatively similar. Japan is way ahead of other countries and the US is a little bit further behind. But, again, broad trends are fairly similar. The idiosyncratic differences within countries are then washed out in a sense that to the extent, as you said, that the global financial markets are integrated. So, in the world of perfect integration of global financial markets, we would only observe one global real interest rate and the differential impacts of, or the impact of these differential trends, in each of these economies would only show up in the quantities, i.e. in the current account surpluses or deficits, and the net foreign asset positions. In other words, there'll be one common global rate. And, at that rate, some countries will find it optimal to borrow, and some countries will find it optimal to lend, depending on whether they're sort of autarkic or a closed economy, hypothetical r* that would be higher or lower than this common global component. And, I think, in the run up to the global financial crisis, the global financial integration has progressed quite a lot. And we have seen a very strong movement across countries in terms of long-term real interest rates. There has been a little bit more dispersion since the global financial crisis and in the run up to the to the pandemic, although the common trend is still very visible. Overall, this integration is much deeper than in the 60s and 70s and so we would expect this trend to be to be quite persistent, unless there's some very, very strong kind of financial decoupling, which I don't think we would expect at this point.

 

Luke Bartholomew 13:16

So just to confront, I think, at this point, a question that some of our listeners might be thinking about when we've described this concept as unobservable - how is it that we might be measuring this unobservable thing that we're talking about? One of the deputy governors at the Bank of England, Ben Broadbent, likes to say sometimes, and I think he's quoting another economist, that r* is a bit like faith in the sense that you might be able to see it by its works but you can't observe it directly. So how is it that we might go about measuring and getting at this concept of r*?

 

Lukasz Rachel 13:54

So absolutely, this is an unobservable concept and sometimes there are these voices, I don't think Ben’s is one here, actually, but sometimes people say, well, can you really set the policy based on something that is unobservable? So, before we get to the to the measurement challenges, I wanted to confront that criticism because it's interesting. I think, ultimately, we must set policy on the basis of unobservables - at the end of the day, even in a simple product market equilibrium with the supply curve and the demand curve, these curves are unobservable and we only observe equilibrium outcomes of quantity and price. But if we want to understand the market, we really need to estimate or figure out what the slopes of the supply or the demand curves are in order to carry out any kind of policy intervention.

 

Similarly, at the macro level the central bank must think about not just what the GDP numbers that the ONS produces are but also what they imply for the output gap, which is an unobservable quantity, because we do not observe the potential supply of the economy. But it's very important for the central bank to get at that in order to project what the ONS’ numbers actually mean for inflation down the line. So, there's no escaping that policymaking requires thinking about models and estimating unobservable quantities. Now, the question is whether r* is..how do we estimate it and whether it is difficult or not to do so. And I should say that it is a difficult concept in the sense that the error bands around any estimates of r* we have are very large. And so our confidence around any specific estimate should be..we should use many models, we should use many approaches in order to get at the most robust answer to this question of where r* is, but I think, in the broad sense of having interest rates on average in real terms being around 4% 25 years ago or so, and the economy doing relatively fine and then having real interest rates basically close to zero or negative in the run up to the pandemic, and the economy not doing so great. That sort of juxtaposition of variables does give us a strong hint that something fundamentally has changed in the economy. And in terms of how we go about estimating those quantities, there are some econometric models which basically take the data that I just mentioned and try to uncover these unobservable components using some filtering techniques. So that's one class of models that one can use. The other class of models that we can use is to write down the structural model of the economy, where there's households and firms and the government and these agents take optimal decisions. And we can then think about what is their desire to save and what is the desire to invest? Given these underlying structural changes in the economy, like demographics or changing productivity growth, I think those are the two broad classes of models that people use to get to this concept.

 

Luke Bartholomew 17:27

So, Bob, I think this would be a good time to bring you into the conversation because in the same way that the fall in real interest rates, combined with a period of economic weakness pre-pandemic, was prima facie evidence in the way Lukasz described it there of r* having fallen, you might say that the big increase in inflation during the pandemic at a time where interest rates stay extremely low is prima facie evidence that r* increased significantly. And without policy rates going up that meant that policy got significantly easier and that's what allowed inflation to become so high.

 

Now, the New York Fed used to publish, and I think has just started publishing again, its measure of r* but famously it stopped doing so during the pandemic because it was going a bit haywire. But you have managed to reconstruct what you think might have occurred to r* during that period. So, can you talk us through what your research suggests r* was doing and why it might have been doing so during the pandemic?

 

Bob Gilhooly 18:40

Yeah, thanks for the call. I'll try not to go kind of too techie on this one, I don’t think people honestly want to hear the minutiae details, but effectively, kind of altering the  Holston-Laubach-Williams model to try to kind of create a short run measure of real equilibrium interest rates. For the US, this shows r* in my estimates was rising from around zero, pre pandemic, to about 2% by the start of 2021. Whereas, of course, the Fed was on hold until early 2022. So, I think this sudden gap, at least in retrospect, helps explain why inflation might have got so out of control. Now, of course, this is not easy to consider in real time but one can point to can a combination of factors which could explain why r* got pushed so sharply higher by the pandemic. Most of this boils down to a policy stimulus hitting into and then amplifying pandemic induced supply damage. So, restrictions designed to protect the population from COVID simultaneously kind of damaged supply, curtailed the opportunity to consume services - you can't go out and do that when you're kind of locked in your house. And at the same time fiscal transfers reinforced household balance sheets, particularly in the US, contributing to labour shortages and then this all added pressure to supply chains as consumption pivoted quite sharply away from services and towards goods. Potentially even you know pressures could have been made worse, by maybe the desire to consume might have actually been kind of rising over this period, a sort of ‘you only live once’ shock, if you will, the discount rate might have kind of been brought about by everyone being forced a little bit to reckon with their own mortality and health prospects. And then you'd also appeal here to your monthly policy reaction being a contributing factor too. The Fed’s flexible average inflation targeting framework (FAIT) was itself born out of a kind of world of low r* that existed pre-pandemic, and with this kind of world of subpar inflation following on from the global financial crisis, you add that to risk management considerations in central banking, which tends to lead policymakers towards more slowly withdrawing support, and you kind of set the stage, if you will, up for a policy mistake. So, in that sense, while  fiscal policy might have been kind of setting the stage here, the Fed not taking away the punchbowl at the most acute phase of the pandemic I think probably also contributed to the short run measure of r* getting pushed higher, thereby letting inflation out of the bottle as the policy rates only moved up with quite a large delay.

 

Luke Bartholomew 21:30

So, one striking feature of the US economy, at least over the last few months is just how resilient it has been to higher interest rates. So, it could very well be on this argument that r* is, in fact, much higher. And so, this increase in interest rates we've seen isn't nearly as significant a tightening in monetary policy as we might think. And that's just why the  economy has been as strong as it has been. So how convincing do you find that argument? Is it possible that r* is even higher than your current estimates suggest?

 

Bob Gilhooly 22:08

Potentially yes, r* estimates are very uncertain as Lukasz was saying, this is a bit of a peril of dealing with, ultimately, the unobservable. Some people have actually been pointing to the 10-year yield in the US as a signal that the market actually believes that r* might have shifted up kind of more  permanently higher. There's actually some kind of technical market reasons as to why we might doubt that signal. There are also some questions about whether the kind of factors that you'd think about pinning down the long run, which you typically think of as being slow moving, could have suddenly jumped. And at the same time maybe there are some reasons to think that actually short run pandemic dynamics, as I described them, are actually maybe in the process of if not fully reversing, at least fading. Global supply chains have pretty much fully recovered, excess savings in the US seem to be falling alongside some improvement in labour supply so that buffer to consume is falling. Core might not yet be target consistent that it has eased very notably with a kind of lift to economic pain. So, you know, appealing to a kind of fundamental change in inflation generation process which keeps r* really high, might be pretty tough. So yes, there’s some risk of r* being higher, but I'd actually maybe frame it as having reasons to think it could be falling back, even if we're kind of unsure about what the ultimate level is. For what it's worth, our estimates of short run r* have actually fallen back somewhat since 2021. by about a percentage point over the course of 2022. So, I think we can actually square the circle though here, in terms of why so resilient, if r* might have actually been easing which would imply policies actually getting tighter. And I think it'd be to consider the timing, not in terms of when the Fed started actually hiking rates, but how policy rates then compared to their equilibrium counterpart. So one narrative here is that 500 basis points of Fed hikes must have already been absorbed, hence the US having kind of a soft landing, or even as some people have been saying, a no landing. But if you think of this in the terms of when the clock starts, and when actually monetary policy becomes tight, well, in our setup this only actually happened in the sort of second half of 2022, when the FOMC had pushed its policy rates in real terms above r*. So that kind of tightening effect could help explain maybe why the economy is actually so resilient, implying that this kind of long and variable transmission mechanism that we often talk about in monetary policy is kind of still working its way through the system. Maybe also just giving another, a different, reason on why the Fed is seemingly happy to kind of sit with rates where they are.

 

Luke Bartholomew 24:56

And then looking beyond the discussion of you called it short term, r*, there's been a lot of discussion about where interest rates might settle in the long run. And, in particular, an idea that's talked about a lot at the moment is that shifts in demographics might mean that equilibrium interest rates are going to be much higher. Lukasz talks about how demographics were an important factor in pushing r* down for much of the previous 30 years. But the thought is perhaps the sign on that might have flipped, that ageing populations are involved in a lot of saving. And that helps to put downward pressure on r*. But old populations, which is where we are at a global level, tend to dis-save - they run down the savings that they've accumulated during their working life when they were ageing. And, as such, desired savings fall significantly. And that puts upward pressure on r*. So, you've done some work looking specifically at demographic drivers of equilibrium interest rates, do you find this argument at all convincing?

 

Bob Gilhooly 26:13

No, not really to be honest, there's a few things to unpack on this one - I should probably give a bit less of a flippant answer. And you know, demographics I think is a really interesting one. As both you and Lukasz mentioned, there's the kind of savings aspect. And then also r* is closely related to the potential growth rate of the economy. So demographics is effectively affecting both aspects, given the shifting population implications, if you will, for savings and investment side if I simplify a bit here, while the number of workers is like a direct input into potential growth. So we really need to consider how the balance of these two effects works, how they could potentially be operating over different timescales too. So we found that, actually, demographics was pretty important in is driving r* down before the pandemic, but we don't really subscribe to this argument that it is kind of radically shifted, the sign has flipped, or indeed that it's necessarily even going to push r* much higher over the next seven to 10 years. It’s good news in this kind of relative sea of uncertainty that demographic trends over the next 10 years or so are actually relatively firmly set, reflecting past trends and fertility. So, we can actually have, within economics, reasonable confidence on the outlook for population, we can use that to help inform our view of potential growth by the implications for the workforce too.

 

After modelling both the global growth outlook and then thinking about how demographic composition might be affecting savings and investments, we can effectively put this all together to consider how those net balances could play out. And we find that the negative impact from potential growth wins out in the end. Even in the case of Japan, which is  well known for its ageing population, we estimate then a more adverse demographic backdrop, is still likely to be putting modest downward pressure on r*. It is actually only much later, beyond 2035 or so, that these kind of indirect ageing effects we think are sufficiently strong to push r* higher. Now, for some countries, such as China and Thailand, we do think they’re  going to have some more upward pressure from demographics on r* earlier on. These countries have very adverse moves in dependency ratios and ageing speeds that are very pronounced, such that the compositional effect can dominate the labour force effect by potential growth. But even then, we actually think that kind of other weakening drivers of potential growth, such as productivity and capital deepening for China, will still on the whole more than offset this. And since we actually find little upward pressure on r* for the major economies, and then we build in this mechanism which allows r* to percolate through the global financial system, we find this global channel should be helping to offset upward pressure from demographics from where it does emerge.

 

Luke Bartholomew 29:17

So, Lukasz we're just bringing you in to the conversation again, I'm interested in where you think r* might settle and just to expand the list of potential drivers on r* And those that are often invoked as ideas as to why it might have moved higher. People talk about changing patterns of globalisation, perhaps fiscal policy has become more active in demand stabilisation and combined with a more active state in investing in green technologies or for other geopolitical reasons means sustainably higher debt to GDP ratio or perhaps AI is going to cause a big increase in potential growth and demand for capital and investment as a consequence of that. So, to ask you the same question I asked Bob, do you find any of that convincing?

 

Lukasz Rachel 30:12

Yeah. So first maybe, let me start by going back to the demographic story that Bob just outlined very nicely. I just wanted to say that I absolutely agree, there is this element of change in terms of the direct effect of the saving propensity of the older population. But that is very much, in the models that we have, this is very much balanced out by basically smaller workforce and slower potential growth. I very much agree that there is no sudden turnaround round the corner coming from demographics, at least in the scenarios that I have studied. Now, in terms of the other factors so, absolutely, there's been some big changes in the last few years in the global economy. One massive change was of course the rise in government debt to GDP ratios during the pandemic. And the question is what that will mean for r*. This increase looked a lot bigger a couple of years ago though, because the last two years of inflation have essentially meant that these government debt to GDP ratios across the OECD have declined substantially. And so that increase in government debt to GDP ratios is probably not as large as we might have thought and could add, perhaps, maybe 20 or 30 basis points to advanced economy r* over the next few years, but that is nothing that would be reversing this long run trend. But I must say that, in the US at least, the government deficit appears to have ballooned this year, I think it's about 2 trillion right now, what people estimate for this fiscal year, and so it remains to be seen how much more the debt grows basically over the near future. And that's something to watch.

 

The other thing that has potentially changed is the attitude towards public investment and industrial policy, particularly in the US but perhaps maybe also in some other countries. There are some early indicators, early analysis or early estimates of the macroeconomic impact of these policies. And they suggest large impact on specific investment within specific areas, like clean energy or semiconductors. But once one brings these numbers to the whole macroeconomy, the numbers look relatively contained. But I think that this area definitely looks to me like it could be the source of upside risk. The reason why it's so hard to assess this with any certainty is that we haven't seen this kind of policies for a very long time, if ever. And so it's very difficult to judge how much of crowding in of private investment the public initiatives will bring. One thing that might be offsetting the impact of the higher investment on r* is that as technology progresses within, say, the green sector, this investment is also going to become cheaper over time and cheaper capital goods have also been one contributing factor to low rates, we basically don't need that much saving to finance investment because investment goods are becoming cheaper over time. And so that's a useful offset to think about. And then, finally, you mentioned AI type of innovations and I think that's another source of risk. The estimates of AI are extremely uncertain at this point and as far as I can tell, they basically cover almost the universe of numbers between essentially zero to kind of three or four % impact on productivity growth every year. So obviously, if you think that AI could boost productivity by several percentage points per year, then we're in a very different ballgame altogether in terms of macroeconomic outlook and r* will be affected as well. But we should also remember the other side of the story. So, the revolution that comes with AI might come with a lot of uncertainty for individual firms, for individual people that could spur some precautionary saving. And perhaps we might be worried about what the balance of market power is going to look like after these technologies settle in our economy and that sort of market power issues that I think we have especially in the US struggled with over the past 20 years or so, they could also play an important role and possibly depress returns on the safe assets further. So, I think overall taking it all together, when I run my models and my analysis, what I tend to think is that the market observed real interest rates just in the run up to the pandemic, say from 2017 to 2020, have declined a little bit below what we can account for in terms of the structural models. And so the recent correction in the market sort of brings us closer in line with what the structural models can predict based on with these different factors that we have discussed. So, I'm not seeing any rapid turnaround from any of these specific factors. Albeit there are definitely some upside risks out there. Instead, what I think might have happened is that the market might have overshot before the pandemic and corrected now to roughly the level where these which we can kind of justify or account for with some of these structural approaches.

 

Luke Bartholomew 36:20

And then finally, Lukasz, I just want to touch on the interrelationship between r* and monetary policy regime. So less the question of interest rates right now that sort of the appropriate context for setting interest rates and the target that monetary policy should be trying to achieve. Because it was a very popular idea pre pandemic, that if r* was going to be extremely low, that meant that policy interest rates on average will be very low. And so, when a negative shock came along, it would be very easy for interest rates to find themselves stuck at the effective lower bound, there wouldn't be much space to cut before they got to something like zero and then couldn't be pushed any lower. And that would lead to monetary policy being, on average, not accommodative enough and so, therefore, perhaps we needed a policy regime that favoured a higher inflation target, or had some sort of memory in the system that required catch up or a different kind of fiscal monetary policy mix. So given the development, since the pandemic, both around r*, but I guess, around inflation too, so how do you assess those arguments now?

 

Lukasz Rachel 37:30

So, given my assessment of where long term r* is ultimately going to settle, I think a lot of these arguments still make sense. To the extent that r* settles at around half or 1%, in real terms in the longer run, that is in my opinion still uncomfortably close to the zero lower bound. And I think the costs of hitting the zero lower bound are substantial. And so, I think it would be worth to revisit the debate about the appropriate inflation target in the long run. This is probably not the debate to have right now given that we are clearly not at steady state at the moment and we have to worry about bringing inflation back to target. But I think that's something to think about going forward. Of course, if real interest rates will settle at a higher rate that will make fiscal policy more expensive. And in particular the sustainable deficit or this fiscal sustainability picture is, of course, very much complicated by higher real rates. And I think that policymakers should worry about the fact that, on the one hand, ageing populations will exert a pressure on spending through social security programmes while, at the same time, the tax base is going to be getting smaller. And all at the same time it could be that r* is higher, and that it is rising even if we don't have any deficit, if we even run a balanced budget, and that's what will happen if R is greater than G. So, I think that is definitely one risk that policymakers should think about in longer term. But, for the moment I think about it as a risk and my central expectation is that still hopefully G is going to be greater than R in the long term once the post pandemic dust settles.

 

Luke Bartholomew 39:45

And that is all we have time for this week. So as ever, please do let me ask you to subscribe and rate us on your podcast platform of choice and then all that remains is for me to thank Bob and Lukasz for joining us today and their fantastic contributions and thank you all for listening. So, thanks very much and speak again soon.