Josh Hendrickson on Using Monetary Policy as a Jobs Guarantee

Josh Hendrickson is an associate professor of economics at the University of Mississippi, where he specializes in monetary economics. He also writes for his blog, The Everyday Economist. Josh is a returning guest to the show, and he joins today to talk about his new paper, *Monetary Policy as a Jobs Guarantee*. David and Josh discuss how monetary policy can be better outsourced to the market as well as the Fed’s past mistakes, and what it can do to improve in the future.

Read the full episode transcript:

Note: While transcripts are lightly edited, they are not rigorously proofed for accuracy. If you notice an error, please reach out to [email protected].

David Beckworth: Josh, welcome back to the show.

Josh Hendrickson: Thanks for having me.

Beckworth: Well, it's good to have you on, and full disclosure to our listeners, Josh and I are co-authors. We write together a lot. We share a similar vision, largely. We have some disagreements of course, but Josh has written another interesting paper, and the title as I mentioned before is Monetary Policy as a Jobs Guarantee and it's a timely piece because there's been a lot of discussion, a lot of interest lately in a federal jobs guarantee program. And, this piece that you wrote, Josh, has come out as a way to maybe help us better understand what a true job guarantee program would be like, but can you go ahead and tell us? Give us kind of the executive summary of your piece and we'll get into it in more detail.

Fiscal vs Monetary Jobs Guarantee

Hendrickson: So, I think that when we talk about the fiscal jobs guarantee, I guess that it's gotten a lot of press lately. Typically, what they're talking about is, they're talking about a proposal where you essentially promise that the government will hire anybody who wants a job at some fixed, real wage. And so essentially, it's a plan to eliminate involuntary unemployment. So, anybody who would be willing to work for that real wage would be able to get a job from the government, and then they can keep that job or transition to the private sector or whatever. But, the key point there is that it does guarantee people a job, but effectively what it's doing is it's just a government guarantee that you can work for that real wage if you are willing to do so.

Hendrickson: So, the monetary policy version of this is somewhat similar. So effectively, if you wanted monetary policy to do this, you could just have monetary policy set up, so that the central bank is willing to buy and sell labor at some fixed price, and we can kind of talk about the mechanics of that, but if you were able to have a monetary policy that does that, then it's effectively doing the same sort of thing as the fiscals job guarantee.

Beckworth: So, it's a very interesting proposal, and if you're not careful, one might read it as very status, very interventionist. It's a job guarantee, but as you actually show in the paper, it's very market friendly and relies on the market. And, you compare it to the gold standard where the government back in the day, when they had the gold standard, would buy and sell gold at a fixed exchange rate or at a certain price, and you argue that the Federal Reserve could do the same thing.

Beckworth: So for example, you mention in your paper if you define the dollar being equal to four minutes of labor, that'd be $15 for an hour of labor. And so, the Fed would stand ready to buy and sell labor at $15 an hour. Again, on the one hand this might sound very interventionist, very state heavy since the central bank would be guaranteeing to buy or sell labor at that rate, but on the other hand you actually argue that this would be a market driven process. Now, walk us through this in more detail, and also tell us, who were the original authors of this idea because you actually talk about that as well in the paper.

Hendrickson: So, this idea is originally due to Earl Thompson, who was an economist at UCLA, and then one of Thompson's students, David Glasner who has told me that he learned about this from Earl, but I think he's being a little bit modest because I think he does a really good job outlining exactly how this would work in his book on free banking. So, I guess, one of the ways to think about this is the reason I draw the analogy to the gold standard is that under a gold standard, essentially what you have is a ... you have the dollar or whatever your unit of account is, you have it defined as a particular quantity of gold. So, if you define one dollar as say 1/20th of an ounce of gold, then that necessarily implies that the price of gold is fixed and is $20.

Hendrickson: And so, what this means is that if you think about supply and demand, when we talk about supply and demand, we often talk about supply and demand as determining the price of a particular good, but what we really mean is it's determining the relative price of that good because we're holding all of their prices constant when we think about a supply and demand. And so, when you think about the gold standard, there's still a supply and demand for gold in the economy, but the nominal price is fixed by that definition of the dollar. So, since that nominal price is fixed by the definition of the dollar, whenever supply and demand change, the real price of gold has to adjust. Now, since the nominal price of gold cannot adjust because of the definition of the dollar, then all other prices have to adjust.

Hendrickson: So, the price level has to kind of adjust when there are fluctuations in the gold market. And in fact, this is kind of one of the major criticisms of the gold standard is that you have this sector, the gold sector, and this is a very small part of an economy. Some economies might not even have gold mining or anything like that, but fluctuations that are going on in that market sort of internationally are causing fluctuations in the price level and in your economy. And so, one of the criticisms is why do we let this really small, unimportant market kind of create these fluctuations in the economy that we could avoid under a different kind of system? And so, what I do is I frame it in this light because I think that this is one way to understand this so called labor standard is in that same sort of context because if you were to fix the nominal wage, then now what's going to happen is that the supply and demand for labor are going to fluctuate.

Hendrickson: But, since the nominal wage is fixed, then the price level is going to have to adjust to these changes in the supply and demand for labor and since monetary policy makers seem to think the labor market as being of primary importance, then it's not going to be subject to the same sort of criticism as the gold standard would be.

Beckworth: Now, I want to get into the operational details in a few minutes, but just a question about the gold standard to really flesh this out. Is it right that the quantity theory does not apply to the gold standard?

Hendrickson: So, there's some debate about this, but what I would essentially say is that really the traditional quantity theory is talking about the supply and demand essentially for money, dollars. So in a sense, the quantity theory holds under the gold standard, but it's the supply and demand of gold that matters, not the supply and demand of necessarily paper that matters.

Beckworth: If you define money as gold, then the quantity theory would still hold. It's just if you think about the actual money, the inside money wouldn't be the case. Now, would that have any bearing then on this system of yours when you think about the price level of determination in it? What would actually ... I guess my bigger question is, what would actually drive the price level in your labor of standard system?

Hendrickson: So, it should only be driven by real factors. So, things that affect the demand for labor, so over the long run, really the only thing that should affect the price level would be productivity. So, you should expect that prices would kind of fall in conjunction with rising productivity.

Beckworth: So, would there still be a nominal anchor?

Hendrickson: So in this case, a nominal anchor would be the nominal wage.

Beckworth: So, prices defined as nominal wages would be anchored. You wouldn't have rapidly accelerating wage growth, in fact it would be fixed by definition it'd be the real wage that would adjust and again, using the analogy of the gold standard, the central bank would be ready to buy and sell fixed amount of labor for just a dollar. Walk us through the actual operational details of this because it's an indirect ... it's not that they can actually do this directly because there's spot market for labor. Labor as we know is very heterogeneous. Your labor, my labor is very different. So, how would they actually do this? Because that is the difference between what you're proposing and an actual gold standard, right?

Operational Details of a Monetary Policy Jobs Guarantee

Hendrickson: Right, so under a gold standard the central bank, or the banking system, or however your institutional structure is set up, essentially there's this willingness to buy and sell gold at a fixed nominal price. So, the labor standard essentially takes that idea and says, well, what if the central bank was willing to buy and sell labor at a fixed nominal wage? Now in reality, they can't do this, and they can't do this because labor is heterogeneous, labor is not sold in spot markets, and so they can't literally buy and sell labor at a fixed price. But, they can do this through indirect convertibility, so you could choose some other asset and you could basically ensure that that other asset is going to buy a fixed quantity of labor. So, in the paper, I give the example of gold. It doesn't have to be gold, it can be any sort of asset that you want to purchase, but with that you can take that commodity and that commodity can be used for indirect convertibility.

Hendrickson: So essentially, the way that this would work is let's say that we use gold, so if you wanted to stabilize the average wage in the economy. So, say the average age of the economy is $15 an hour and the price of gold is $1,500, then what this would mean is that an ounce of gold would buy you 100 hours of labor. And so, the basic idea here is that what I could do is I could go to the central bank and the central bank would basically take this asset and they'd be willing to buy and sell that asset at whatever the market price of that asset is, but what they would do is they would guarantee that the nominal value of that asset would pay for a particular quantity of labor.

Hendrickson: So in that example, if I could go to the central bank and I could buy an ounce of gold for $1,500, that should buy me 100 hours of labor if the wage is $15. However, if it comes out that the actual wage is really, say 16.50 and hour, then I'm not getting 100 hours of labor. In fact, the price of labor is 10% higher than I expected. So, the way that this would work is that the central bank would basically collect fees from anybody who had sold gold in this scenario and they would pay a rebate to anybody who would purchase gold from them. So, if you had purchased gold for $1,500 an ounce and it turned out that the average wage was 16.50 and not $15, then you would have actually needed $1,650 to buy a 100 hours of labor. And so, they would effectively owe you $150. So, anybody who bought gold in this scenario would get $150 rebate from the central bank, anybody who sold gold during that month would have to pay $150 back to the bank.

Hendrickson:  So, this sounds kind of complex, but the way to think about this is suppose that everybody thinks that nominal wages are actually going to be higher than the target. So, if the nominal wage index is going to be higher than the target, than people aren't going to want to go buy gold because you can make an arbitrage profit. So in other words, you could go to the central bank, and buy gold from the central bank at $1,500 an ounce, turn around and sell that on the open market for $1,500 an ounce. So, your investment is $0. But, if wages end up being higher than you anticipated like in our example, you would get this $150 rebate. So in other words, you're not putting forth any sort of initial investment, but you're getting $150 for being correct.

Hendrickson: So in other words, it's an arbitrage profit. It's pure arbitrage, and so the idea here is, is that if you expect wages to go up in the future, or if you expect wages to be higher this month than the target or whatever, then what you're going to do is you're going to be buying gold from the central bank, and then trying to earn that arbitrage profit. And so, if you turn out to be correct, you're going to earn that arbitrage profit, but if you think about what's going on in the background is this is essentially kind of like a futures targeting thing that Scott Sumner talks about. So, what you're really doing here is by predicting that wages are going to be higher than the target. Everybody is going to have an incentive, who has that belief, is going to have an incentive to go to the central bank and purchase gold, but in doing so, what the central bank is doing is it's conducting an open market sale. It's selling gold and it's taking the dollars away.

Hendrickson: So effectively, what's happening is that they're conducting open market sales at precisely the time when people think that wages are going to be higher than anticipated. And so, that should push wages back down towards the target. So in other words, whenever anybody thinks that nominal wages are going to rise, then we should expect that the central bank will be conducting open market sales and those open market sales should then dissipate those expectations and get us back towards target. So effectively, it kind of has that feature that Sumner's proposal has that you're kind of outsourcing monetary policy to the market and not just to the FOMC.

Beckworth: So again, the idea is if we are traders or if we're people in the market place, and we think nominal wages are going to be growing faster than normal and put the nominal wage above the Fed's targeted exchange rate between labor and gold, if gold is the asset you're using, then we are going to rush in. We're going to be the ones who instigate this action and the Fed's going to be kind of passively responding to us. It's going to passively sell us this contract and some if it's we're the catalyst and the Fed's going to be the party that responds to us based on our expectation of the future labor market. Is that right?

Hendrickson: Right.

Beckworth: So, if I'm Jay Powell and I'm sitting in my office at the board of governors, I'm looking at daily the market for gold and some nominal wage index, is that what I'm keeping my eyes on?

Hendrickson: So effectively, the reason that you would use indirect convertibility is for a couple of reasons. So number one, you can't buy labor on the spot market, but number two, we also tend to get the information with a lat. So, if I want to know what average wages are in January, I'm not going to find out until February. So essentially, what you would have the central bank do is their primary responsibility would just to be willing to buy and sell whatever this asset is like gold for the current market price. And, people's expectations about what nominal wages are going to be this month are going to determine whether the Federal Reserve is conducting open market purchases, or open market sales.

Beckworth: All right, so in other words, Jay Powell doesn't even need to sit at his desk, he can go home. He could have a computer basically respond to the offers of people coming to the Fed. Is that right?

Hendrickson: Yeah, so essentially, you can just outsource monetary policy to the market.

Beckworth: Now, this proposal is similar to Scott Sumner's proposal. Now in his case, he's looking at nominal GDP, you're looking at labor income. They tend to be correlated, very similar, but they are different. I guess the second similarity between your proposals, he would have a nominal GDP futures contract where you would just have some kind of asset, right?

Hendrickson: So, in a lot of ways, this proposal acts just like a futures market would act, but we're just using indirect convertibility because of that lag in when you find out the information.

Beckworth: Now ... go ahead.

Hendrickson: What I would say is, I think one of the advantages of this sort of proposal is that under this proposal one of the criticisms that Scott has received is that, well, we don't have nominal GDP in futures markets historically. And so, maybe there's just no interest in it, or maybe the market wouldn't have enough participation. There's all kinds of criticisms there basically related to the futures element of it. And so, I think one of the advantages of this proposal is that you can choose an asset that is traded in international markets where there's enough trading that's going on there regardless of what's going on with monetary policy that you're going to be able to engage in this sort of very liquid market to begin with. And then, you don't have to worry about if we create this market, will people participate, and that sort of thing.

Beckworth: So, you just want an asset, it doesn't need to be gold, that's highly liquid, that has lots of volume, and that's easy for the Fed to use in the monetary policy. And, this leads me to another question, a related question. There's been other proposals to target nominal wages. For example, Greg Mankiw and Ricardo Reis in 2003 had a paper where they're looking at what is the best index to target. The title of the paper is What Measure of Inflation Should a Central Bank Target? Let me read a little excerpt from their paper. What they do is they go after a stability price index. What price index would be the most stabilizing to the macro economy? And they say, "A stability price index depends on the sector's characteristics including, size, cyclical sensitivity, sluggishness of price adjustment, magnitude of sectoral shocks. When a numerical illustration of the problem is calibrated to U.S data, one tentative conclusion is that a central bank thatwants to achieve maximum stability of activity should use a price index that gives substantial weight to the level of nominal wages."

Beckworth: So, they arrive at a nominal wage target as the outcome of an exercise where they're trying to create an index that'll bring economic stability. So, they end up in a similar direction, although there's a big difference. I think the big difference is your approach would make it automatic. The market itself would be effectively doing the monetary policy. They would be the ones coming to the Fed at that fixed price. Wherein the Mankiw Reis approach, it would be still discretionary, it would still be the central bank targeting index. In both cases, there would be a nominal wage index that would be looked at and be examined a target, but in your case, it's more automatic. In their case it's more discretionary, is that right?

Hendrickson: Yeah, so I think that's the primary difference is that they would just ... their kind of idea I think would just be to have the central bank stop looking at, say the inflation rate and start looking at the wage index. I think that's the primary difference in that here what's happening is that you're actually sort of outsourcing monetary policy to market participants.

Beckworth: Now, they motivate this by arguing that it's important to look at nominal wages. They're sticky, they're sluggish, what is your motivation for going to wages versus some other measure?

Using Wages vs. Other Macroeconomic Measures

Hendrickson: So, I don't wish to dispute what they're arguing. I think that they're making arguments that are good arguments in favor of wage targeting in general, regardless of how you do it. But, I also would add, if we go back and we draw this analogy with the gold standard is that you're forcing, under a gold standard, you were forcing all these other markets to adjust to what's going on in the gold market. And so here, you're kind of forcing ... when you target nominal wages, you're forcing every other market to kind of adjust to what's going on in the labor market. And so, if you think about the Fed's mission, one of the Fed's mission is to achieve maximum employment. And so, if you are trying to achieve full employment, if the labor market is of primary importance to you, then you might be willing to bear the cost of having all these other markets adjust to what's going on in the labor market rather than the alternative.

Hendrickson: Ideally, under a wage targeting regime, what you have is you would eliminate any sort of nominal sources of fluctuations in the labor market, which is pretty much the only thing that we can ask a central bank to do if we want them to promote full employment.

Beckworth: So, you are also interested in stabilizing a nominal wage index. You do it different ways in many of the popular calls today. So, I mentioned Mankiw and Reis, but other people have also called for the Fed to target something like a nominal wage index. Now, both of these measures I mentioned earlier are similar in spirit to a nominal GDP target, or a nominal income target. Do you have any preference of one over the other?

Hendrickson: I don't know that I have a preference of one of the other. I would prefer either one to what we currently have, and I guess that's the way to say it.

Beckworth: I wonder though, thinking about this. A nominal wage index target seems like to me might be more robust for different situations in the following. Since you're a small, open economy, and your exported in a sea of a commodity, the normal GDP volatile if exports swing wildly for exchange rates. So, if you had something like a nominal wage index, I think that would be something to be easier and more manageable to target than nominal GDP itself. And therefore, it'd be an approach that's more robust across different situations, different countries, but with that said, I think it's important to me we spend a little more time on this. So, whether you do the labor income target, the nominal income target, let's transition and think about that in terms of a monetary policy rule or a reaction function, like a Taylor rule.

Beckworth: So, a Taylor rule, and so all our listeners know, Taylor's the central bank to adjust the target interest rate, or the tool the Fed typically uses in responses to changes in the output gap, the capacity of the economy and also in response to changes in inflation deviating from its target. How would you see a rule or a reaction function set up that would go from a Fed's instrument all the way to its target? And here in this case, the target presumably would be a nominal wage index or a nominal GDP targeting. How do you see that working?

Implementing a New Monetary Policy Rule

Hendrickson: Well, I would prefer that we talk about monetary policy in terms of money, but I'm afraid that's going to take this conversation in a different direction, but what I would say is I don't think that a nominal wage regime would necessarily ... if we're not going to adopt my proposal, if we're just going to tell the central bank they target nominal wages, I think that it doesn't have to deviate that much from what it currently does. So, you would just want to have the central bank react to these deviations of nominal wages rather than trying to react to deviations of inflation from target and deviations of output around potential. I think that the benefit of switching to a nominal wage regime is if you think about the Taylor rule and you think about the output gap, there are a lot of reasons why we can have an output gap.

Hendrickson: So, maybe a real GDP is changing, maybe a potential GDP is changing, maybe both are changing. And so, you sort of have to navigate what's going on with real GDP and potential GDP in real time, and I just think that's too hard of a problem to respond to in real time. And so, if you're responding to something like the nominal wage, then what this means is that you're going to get price stability in terms of wages, and you're going to eliminate any sort of nominal sources of these fluctuations in the labor market. And so, what you should do is by minimizing these nominal deviations, you should take care of that output gap issue without actually having to look at the output gap.

Beckworth: That's the knowledge problem in monetary policy. Now, you made an interesting point. You would rather think of doing monetary policy in terms of money, so let's talk about that. I think that's an interesting conversation. I had David Andolfatto on the show previously not too long ago. We talked about the inability of Phillips curves to be very helpful currently in monetary policy. Maybe one could argue never were that helpful, but currently, FOMC officials are trying to find what's going on with inflation, they're looking at their Phillips curve, they're looking at the low unemployment rate, and they're a little puzzled that they've said so themselves.

Beckworth: And so, David Andolfatto had some posts, made some comments, he got some blowback from Paul Krugman, from others because what David said is he criticized the whole Phillips curve approach that looks at inflation as a function of output gaps, capacity, and he suggested instead we look at inflation and more broadly speaking, monetary policy through the lens of money as you've just mentioned. And, he said, "You need to look at money, supply money, demand." And he said, "Yeah, we need to be more broad minded about it." Money supply would include things like treasury bills, commercial paper institutional money has. It's not just the more narrow money assets. But he says once we do that, he says we have a more practical framework from which to think about monetary policy and inflation. What are your thoughts on that?

Hendrickson: So, I think I'm motivated by the fact that a lot of the criticism of using money as a guide for monetary policy is kind of based on flawed empirical analysis. So, you'll notice a lot of times that people will say things like, "Well, we can't use monetary aggregates to tell us anything about the stance of monetary policy because we've got these studies over here that show that the standards are of old monetary theory predictions don't hold when we look at monetary aggregates using modern data and things like that." And, I think as we've kind of talked about before, the problem is, is that these all rely on simple sum aggregates and the broader that you get when you're aggregating across these different types of money assets, the more of a problem that you're creating in terms of aggregation.

Hendrickson: The way that historically, monetary aggregates were constructed, it's just by adding up all of these different types of money assets, and the problem with that is that in terms of index number theory, in terms of aggregation theory, what that implies is that all of these assets are perfect substitutes. Now, we know in reality that they're not perfect substitutes, and so in theory this form of aggregation is wrong, and it turns out that in practice it's also wrong. So, there are a number of papers now that have shown that if you use Divisia monetary aggregates, for example instead of simple sum aggregates, these Divisia aggregates are derived from micro economic theory, and so if you use these aggregates instead of the simple sum aggregates, what you actually find is that the traditional relationships between money and inflation, money and nominal income, these sorts of things are actually there in the data.

Hendrickson: It's just that when you're using these simple sum measures, you're using a very flawed measure of the money supply and that's causing people to say, "Well, money's not entirely useful." But, if you use these Divisia aggregates, then you get the predicted relationships that any sort of quantity theorist would understand.

Beckworth: So in other words, real money demand is still stable, is that right?

Hendrickson: Right, so money demand is relatively stable over a long period of time. And so, these aggregates can be a useful indicator of the stance of monetary policy. I know you had Peter Ireland before and he has a paper with my colleague Mike Belongia, and they go through how would you actually implement nominal GDP targeting and they go through and actually discuss how you would do it using a variation of what's called the P-Star model.

Hendrickson: So, that was a model that was used to determine how much money you would need in circulation to achieve your price level target and things like that, and they basically go through and show that when you use these Divisia aggregates, you can adjust the monetary base and if you have some knowledge of what the money multiplier is associated with these particular aggregates, that you can actually forecast these things pretty well. And so, if you can forecast nominal GDP pretty well just by knowing the base and the multiplier, then what that suggests is that those broad monetary aggregates are sort of useful intermediate targets for figuring out what's going to happen to nominal GDP.

Beckworth: Yes, we also had Bill Barnett on the show and he also walked us through that discussion, how these measures of money, these Divisia measures are very useful and show a stable real money demand relation over longer periods. But kind of going back to your proposal, which again, it ultimately is targeting stable nominal wages or Scott's proposals of targeting nominal GDP. Let's say the Fed moves in that direction. Let's say the Fed says, "Sure, we're interested. We like your arguments." But, they still do monetary policy through discretion, through FOMC meetings, the way that it's done and there's a target to something you like.

Beckworth: What kind of policy rule would you recommend for them? In other words, what kind of Taylor rule would you prescribe for them? Now, they don't religiously follow the Taylor rule per se. In fact, they have a list of rules. Under the Jay Powell Fed, they've increasingly been publishing a variety and many of choices of monetary policy rules. To be fair, it's started under Janet Yellen, but Jay Powell has been very eager in his endorsement of these rules, but they're more of a guide. So, if you were going to advise Jay Powell and the Fed and they had adopted a nominal GDP target, and they wanted some kind of reaction function and they came to you, how would it look?

Constructing a New Reaction Function for the Fed

Hendrickson: If I'm going to design a reaction function, my reaction function is probably going to be based on something like the monetary base in conjunction with these multipliers for the broader monetary aggregates. And then, using these monetary aggregates as sort of intermediate targets for the ultimate goal, whether it be nominal wages, or whether it be nominal GDP. To me, I think that part of the problem that we have with current policy kind of goes back to something that I've written about before and really, I have to give credit to Robert Hetzel for kind of originally teaching me this point, is that too much of Fed policy right now is focusing on reacting to things.

Hendrickson: So, if you think about the problem in the 1970s, the problem in the 1970s wasn't that the central bank wasn't responding sufficiently to inflation. The idea that the Fed fights inflation is a weird terminology because the Federal Reserve creates inflation. So, if you say that they fight inflation, that seems to suggest that inflation is coming from somewhere else, that inflation just kind of happens, whether it's these excess capacity stories or whatever, that inflation just sort of happens and the Federal Reserve has to react to it. But, I think we need to get back to a monetary policy where the Federal Reserve is taking credit for its actions. So, what I mean is that the thing that changed when Volcker came in was not that he was a better inflation fighter, it's that when he came in he was explicit about the fact that, look, if inflation is high, it's our fault and we can stop it.

Hendrickson: And so, we sort of have this story now where the Fed is confused as to why we're not seeing more inflation and they keep referring to Phillips curve logic and things like this, but that puts them in sort of the passive role of allowing inflation to happen or fighting inflation when it gets too high. When in reality, the central bank determines what inflation is actually going to be. Now, I don't mean to minimize what they're doing because, look, to say that the central bank can create inflation is a really easy thing to say. To say that they can create 2% inflation next month is a different statement, but the way that they're phrasing things kind of reveals to me that they're kind of going back to this way of thinking about monetary policy that we had in the 70s where the Fed is sort of like a passive player in this where if they want to allow more inflation, then they'll allow more inflation. If they want to fight higher inflation, then they'll fight higher inflation. In reality, they're the ones who are creating inflation, so if they want more inflation, they can create more. If they want less, then they can reduce inflation.

Beckworth: Now just to be clear, are you saying they can create inflation instantly? Are you arguing more they control the medium to long run path of inflation?

Hendrickson: Well, I'm not going to argue in favor of the caricature of the quantity theory where the only thing that matters is the money supply. I'll fully acknowledge that there are other shocks that occur in the economy that have some effect on the price level or some temporary effect on the inflation rate, but yes, over the medium not the long term. It's monetary policy that determines what the inflation rate is. And so, you can have these shocks that are going to push it higher than expected or lower than expected or whatever, but over the medium and long term, the Federal Reserve is determining the rate of inflation.

Beckworth: So then, my criticisms that I have had on the show, one my writings said the Fed is ultimately responsible for the fact that inflation has been below its target for the past decade are reasonable, is that right?

The Fed’s Culpability for Low Inflation

Hendrickson: Yeah, that's correct. And, I want to be careful here because there's a way to take what I'm saying wrong, and so if you have 1.5% inflation and your goal is to get to 2% and you keep producing 1.5% inflation or like a 1.7% inflation rate, I think that what we need to do is think more in terms of kind of a range. Maybe if we want 2% inflation, what that really means is we want something between 1.5 and 2.5%. And then, anything that's kind of there is fine because I don't want people to think that I'm making the criticism that if the inflation rate is 1.5% that they could just make it 1.7% if they really wanted to or something like that. This isn't about how many basis points they can add to ... the inflation rate, it's just a general idea. If you're persistently not hitting your target, then it's your fault that you're not hitting your target, but I do think we need to talk about this within a range because when we're talking about very low rates of inflation, I do think that there could be some issues associated if you have 1.5% inflation and your target is two, how acceptable is that? Is that something that we're willing to deal with?

Hendrickson: Because it is going to be very hard to micromanage the inflation rate because we're using these blunt instruments. We're trying to do open market operations or right now using interest on reserves to try to push things where we want to go. And so, it's not going to be an exact science, so I don't want people to get the impression that that's what I'm saying, but there does need to be this mentality that if you need higher inflation, the central bank is what's going to be the source of that higher inflation. If you want lower inflation, the central bank is going to be the source of that, and too oftentimes I think this is framed in a very passive way that we saw in the 1970s where, well, we're maintaining our stance of policy because we think that GDP is going to be higher and that will lead to higher prices. I think that's the wrong way to think about how monetary policy works.

Beckworth: So, someone who shares your view, Josh, is Bennett McCallum. He's a pretty well-known macro economist. For all you listeners who don't know, look him up, read his work, but he is someone who's shared your view very monetarist in nature, but someone who also is really into DSGE modeling and took macro seriously and he is well-known for proposing an approach to monetary policy called the McCallum rule.

Beckworth: And in that rule, he too would target nominal GDP, but he would do it by having the Fed adjust the growth rate of the monetary base and then he would also be looking at kind of intermediate guides such as the average velocity over the past four years past misses in nominal GDP and so forth. But basically, he would link changes in the monetary base to ultimately changes in nominal GDP. Now, here's my question, would this rule work? I ask this question in light of standard view that was developed first by Bill Poole in the 1970s about whether a central bank should use short term interest rates to conduct its operations, or should it use the monetary base.

Beckworth: And basically, what he concluded that money demand shocks are the key source of disturbances and you use interest rates because it implies you'll offset money demand shocks, and then that's kind of the consensus view. That's why we have central banks around the world today that they use interest rates and their argument is we're going to smooth interest rates out, we're going to offset money demand shocks. So, with that framework in mind, and maybe you want to criticize that framework, could a Fed implement something like a McCallum rule?

The Possibility of a McCallum Rule

Hendrickson: I think they can certainly implement the McCallum rule if they were were interested in implementing it, but I think that one of the reasons why I say I would like them to get back to thinking of this in terms of money rather than interest rates is that interest rates ... I'm not sure how much control that they have over interest rates. There are all sorts of issues here. So for example, some people talk about monetary policy as though there's some underlying natural rate and so what the central bank has to do is to kind of adjust the interest rate to kind of mirror whatever the natural rate is. But, there's a question about what role the central bank plays in this story because sometimes in this story, the central bank is kind of the activist where there's some natural rate that doesn't really change that much and then so when the Federal Reserve or whatever central bank we're talking about adjusts the market rate, then it's deviating that market rate from the natural rate.

Hendrickson: But, there's another view that seems to imply that the natural rate is fluctuating over time, and it's the job of the central bank to adjust policy when that natural rate changes. And so, there's that aspect of the story, and to me that gives you to different ideas about what a central bank is doing and can do, but the other thing is that interest rates in the United States, they're not independent of what's going on elsewhere. So, if you think about all of these ideas about purchasing power parity, about interest rate parity, and things like that, if you look at the gold standard era when we had fixed exchange rates, I think you can find some pretty strong support for these kinds of things. Now, since we've moved to a world without fixed exchange rates in most of the developed world, a lot of people have argued that these sorts of things don't hold or that they take a lot longer to adjust or what have you.

Hendrickson: But on the one hand, most of the things that we say about international trade rely on these characteristics to hold, at least in the long run. And so, if they do hold, then how much control does the central bank actually have over the interest rate? It's not clear to me, and so I would prefer that they conduct policy with the things that they directly control. So, the Fed controls the size of its balance sheet and so if they focus on the monetary base and they use broader monetary aggregates as intermediate targets, I think that that gives us a better way of conducting policy and it also gives us a better way to analyze policy because we can observe what the central bank is doing directly without these external factors that might also be affecting these rates and things like that. I'll give you one example.

Beckworth: Sure.

Hendrickson: When I talk to economists about forward guidance for example, what they say is, well, we're going to keep interest rates low for a long time and we're going to do that because we want to send the market the signal that we're going to be expansionary longer than we should we effectively. And so, that should engineer an increase in output now because we're essentially promising to be irresponsible at some point in the future. The problem is that when I talk to non-economists about this, what they tell me is, well, when the Fed says they're going to keep interest rates low for an even longer period of time, that tells me that things must be worse than they actually are, or that I actually believe them to be. And so, that gives them the precise opposite conclusion that the Fed wants them.

Hendrickson: Now, I don't know how general realizable these things are, but this is kind of the problem when you're targeting an interest rate is that interest rates move for a variety of reasons. The interest rate is affected by people's time preferences, it's affected by expected inflation. It's also affected by people's expectations of growth in the economy. All of these things are affecting interest rates, and so if all of these things are affecting interest rates, it's hard to look at an interest rates and get some indication about what the stance of policy is or what's going to happen with policy going forward because there are so many different factors that affect the interest rate that it's just hard to determine which factor is operating here and what are they trying to accomplish with policy. We talk about low interest rates to engineer higher inflation, but in the long run if we want higher inflation, we're going to want higher interest rates. And so, there's this difference between the liquidity effect and the Fisher effect and now that you have the Federal Reserve conducting policy by controlling interest on reserves, it's unclear what a liquidity effect even means in that kind of context.

Beckworth: Another reason why I think the monetary approach is also compelling is that there's a political economy angle to this. If you tell people that the Fed is going to hold rates low for an extended period of time, they instantly begin to think of financial repression. The Fed's taking advantage of savers, it's being unfair, and it creates this bad optic. Where if you were using just money, it wouldn't be as much of an ordeal. It wouldn't create the bad optics. If you were using the monetary base, control the price level or nominal GDP, I think you avoid some of the bad optics created from using interest rates.

Hendrickson: I would also point out one more thing, and that is we know substantially less about the monetary transmission mechanism than we should. And so, what I mean by that is we need to better understand how monetary policy actually works and what sort of channels that monetary policy works through, and the kind of implicit argument or sometimes explicit argument when it comes to interest rate targeting is that actually all we need to care about is the interest rate because the interest rate is going to ... when it's these movements in the interest rate that transmit monetary policy, but I think if that's the case, then monetary policy is probably a very weak tool.

Hendrickson: But, if we look at the evidence, it seems like monetary policy can have pretty enormous effects on economic activity and so it's kind of like, how do you reconcile the fact that monetary policy seems to be really important, but the empirical evidence suggests that interest rates don't matter that much when you're looking at things like investment and things like that. And so, the question is, what's the mechanism by which monetary policy works? And, I think a lot of the focus on interest rates has kind of reduced even the interest in trying to figure out how monetary policy works because if you're just adjusting interest rates, then it's just a matter of how interest rates affect behavior. That's all we need to understand, but I think that there's a lot of literature in macro that has looked at the monetary transmission mechanism and it teaches us important lessons about the different sorts of effects of monetary policy on the economy that aren't picked up by the interest rate.

Hendrickson: That's not to say that interest rates aren't important at all, but that they might just be one component of a much larger story, and I think if you look at the research on the monetary transmission mechanism, I think that's the impression that you come away with is that, sure, interest rates matter, but interest rates are only part of the story, and that there are these other factors that we need to think about as well.

Beckworth: What's interesting along those lines is that the Fed itself began stressing these other channels during the past decade with the zero lower bound. So, during the zero lower bound, they started stressing the portfolio balance channel, which stresses that the Fed's transmission, the Fed's effect is felt through a wide spectrum of assets through shrinking of those premiums across many yields, many asset prices, and now that we've returned to non-zero lower bound territory, the focus automatically goes back to the expected path of the target interest rate.

Beckworth: So, it is interesting, they acknowledge this point you're making that there is a broader transmission mechanism, but now they kind of fall back to the default setting once we're out of the zero lower bound. So, any parting wisdom you would leave for the current Federal Reserve? What would you encourage them to do as they wrestle with low inflation, interest on reserves, shrinking its balance sheet, any just practical suggestions you would have for them?

Parting Wisdom for the Federal Reserve

Hendrickson: My suggestion would be to get back to doing monetary policy the way that it's traditionally been done, and so what I mean by that is I think that they need to get rid of the interest on reserves regime that they have. Now, I don't ... part of this goes back to, is the interest rate the only thing that matters, but I also think that as George Selgin and David VanHoose and many others have pointed out, this actually kind of creates a sort of regime shift in banking in terms of bank's preferences and of what they want to hold and things like that. And, I don't think that we really understand enough about this interest on reserves regime to be making policy in that way. And so, I would prefer that when the Fed talks about normalization, I would prefer that normalization doesn't mean getting interest rates back to the historical average. I would prefer that that means getting rid of interest on reserves and getting back to sort of normal central bank balance sheet policy.

Beckworth: All right, well with that, our time is up. Our guest today has been Josh Hendrickson, who's authored a new paper titled Monetary Policies as a Jobs Guarantee. We will have that article linked at the SoundCloud webpage. You can also go look it up online. Josh, thanks for coming on the show.

Hendrickson: Thanks for having me.