IS-LM is not a useful model

I have to admit that I have rarely used the IS-LM model since I became interested in economics. Instead, I find the AD-AS model much more useful. Of course, this does not mean that there is a major theoretical flaw in the IS-LM model. Rather, I think that the inferences that economists make by looking at the model are wrong, and that the model leads both economists and students to “reasoning from a price change.” To be more precise, it leads them to assume that changes in the interest rate are a good indicator of the monetary policy stance.[1]Sumner, Scott. A Critique of Interest Rate–Oriented Monetary Economics. Working Paper, Mercatus Center at George Mason University, November 23rd, … Continue reading But that’s not true. For example, interest rates fell in late 2007 due to the weakening economy, and monetary policy was certainly not expansionary. It is true that a shift in IS could cause this, but I suspect most economists either ignored or failed to see it.

Nor do I find it reasonable to assume that increased savings result in lower demand and output. Whether an increase in hoarding is accommodated by the central bank seems to me a more accurate perspective. [2]Rowe, Nick. The paradox of thrift vs the paradox of hoarding. Worthwile Canadian Initiative, October 01st, 2010. … Continue reading However, when looking at monetary policy through the lens of interest rates, it seems clear that this will hardly be seen. However, when looking at monetary policy through the lens of interest rates, it seems clear that this will hardly be seen. Accordingly, when the propensity to save decreases, the central bank has to reduce the money supply. But the increased desire to save does not affect demand, so there is no reason for an inflation-targeting central bank to reduce the money supply.

Let’s start with interest rates. Theoretically, a doubling of the money supply would almost immediately double the price level. But that’s not the case in the real world – mostly because of sticky wages and prices, although there are other factors. If the price level is not doubling, another variable must double to equate the increasing supply. That variable is the (nominal) interest rates, and they are not sticky – they quickly move the equilibrium rate at which money supply and demand equated. In this example, a fall in the interest rate may be a sign of wage and price stickiness rather than the stance of monetary policy. It need not have an essential role in nominal aggregates.

Someone might say, “But Kursad, the IS-LM model is not about nominal aggregates; it’s about real GDP.” I agree; we can’t get real effects with a perfect wage and price elasticity model. But in the case of wage and price stickiness, we can still get the same kinds of real effects, even if interest rates do not play an essential role: it is possible in the SRAS model. Consequently, a sustained increase in the money supply will proportionally raise the price level, or NGDP. The nominal expense ratio also increases, but to a lesser extent. But I don’t see anything necessary for short-term interest rates to fall in this process. If current rates are held constant, real interest rates will fall as expected inflation increases. Or nominal rates may rise at the same rate as expected inflation, and real rates remain constant. In this case, the opportunity cost of holding cash increases rapidly. We can get real effects when we consider nominal expenditures to rise and wages/prices are sticky. Or suppose nominal rates are falling. I think this would be the case that the IS-LM model would most easily explain. But I don’t understand exactly why it matters whether interest rates fall when the expected future NGDP rises. This puts upward pressure on existing AD and NGDP, and this is true for any commodity. So why are we focusing on interest rates? My conclusion from these scenarios is that monetary policy determines the growth rate in NGDP, which determines the level of nominal interest rates. Interest rates, then, do not determine the stance of monetary policy, because they are not a cause, but rather a result.

But none of this is why I choose not to use IS-LM. Actually, I have purely pragmatic reasons. Basically, I think it’s hard to apply the IS-LM model to the real world, and we have better alternatives instead. I saw this more clearly when I read this passage by Robert King[3]King, Robert G. 1993. “Will the New Keynesian Macroeconomics Resurrect the IS-LM Model?” Journal of Economic Perspectives, 7 (1): 67-82.DOI: 10.1257/jep.7.1.67:

“If changes in the money stock are persistent, then they lead to persistent changes in aggregate demand. With a rational expectations investment function of the neoclassical form, persistent changes in demand for final output lead to quantitatively major shifts in the investment demand schedule at given real interest rates. These effects are generally sufficiently important that real interest rates actually rise with a monetary expansion rather than fall: the IS curve effect outweighs the direct LM curve effect. Further, if persistent changes in the money stock are only gradually translated into price or wage increases, then there are very large additional expected inflation effects on nominal interest rates.”

When I read this, everything actually became clearer. Everything can be interpreted with IS/LM model; still, that doesn’t make it useful. If it’s not intuitively obvious, most people can misunderstand it; and if most people misinterpret it, it’s useless.

While the example I give most of the time relates to the Great Recession, there are many other examples as well. Similar things happened in 1920-21 and 1929-30. In both cases, monetary policy was extremely tight, but if you interpreted it in terms of interest rates, you would think that monetary policy was “appropriate.” Similar situations apply for 1929, 1937, and 2008. However, the IS-LM is not a useful model when thinking about these shocks and seems to mislead most economists.

Between 1926-29, gold reserve rates were increasing at about 2.5% per year. Between October 1929 and 1930, this ratio rose to 9% as central banks increased demand for gold. This naturally changed the future path of NGDP, thus leading to a fall in stock and asset prices. Keynesians saw nominal interest rates fall during this period and assumed that the problem was not monetary policy. I see precisely this kind of reasoning as a bad interpretation of the IS-LM model. Someone might say, “if the problem is not with the model but with the interpreters, you can’t blame the model.” However, I am not blaming IS-LM or claiming it has a theoretical flaw. Rather, I question how useful a model could be that many of the leading economists of the time misinterpreted. If you misdiagnose the disease, your prescription will be useless.

It must be admitted that the IS-LM model works quite well in Great Moderation. However, this seems to me more related to the Taylor Rule. However, the model is not that useful in times when inflation expectations change rapidly. It is more difficult to predict the real interest rate during these periods; For this reason, it is much easier to think that the monetary policy is expansionary in periods such as 1929, 1937, and 2007 when it is actually tight. In a scenario where the Wicksellian equilibrium rate falls more than the policy rate, IS-LM would be misleading. Likewise, traditional monetarism will fail to interpret these periods. Consider a driver who drives well on a straight road, but is more prone to crashes when there are many intersections. This is exactly how I see IS-LM.

All right, so what’s my solution? I prefer to use a more useful and simple model. I prefer to rely on AD/AS, where output changes depending on supply and demand shocks. I believe the interest rate transmission mechanism is confusing and misleading. The AD curve is a rectangular hyperbola (simply NGDP, or nominal expenditures) AS is merely a function that represents short-run output. And finally, OL is a vertical curve that shows the long-run output, i.e., the optimal level of the economy. Since nominal wage stickiness is the underlying assumption of this model, the “Optimal level” will vary depending on the type of wage stickiness and will require us to define nominal shocks in terms of wage stickiness. We could actually add GDP-linked bonds to this model. It would be much easier to model this way, but as we don’t have such a thing for now, we have to make do with what we have.[4]Prasad, Pradyumna. Why you should issue GDP linked bonds, Bretton Goods Substack, July 20th, 2021. https://brettongoods.substack.com/p/why-you-should-issue-gdp-linked-bonds. An unexpected change in AD/NGDP/NE creates a change in hours worked as wages are sticky. The output will returns to OL when wages are fully adjusted. The simple implication of the model is that NGDP shocks affect both production and inflation. I think this model, the “musical chairs”[5]Sumner, Scott. Money and output (The musical chairs model), The Money Illusion Blog, April 06th, 2013. https://www.themoneyillusion.com/money-and-output-the-musical-chairs-model/ interpretation, is much more useful than the interest rate-oriented approach.

P.S. I didn’t originally think of the AD curve as a hyperbola, but as I said here, after seeing Sumner’s model, I decided to tweak it a bit. I think Sumner’s framework has been much more useful.

References

References
1
Sumner, Scott. A Critique of Interest Rate–Oriented Monetary Economics. Working Paper, Mercatus Center at George Mason University, November 23rd, 2020. https://www.mercatus.org/system/files/sumner-critique-interest-rate-mercatus-working-paper-v1.pdf.
2 Rowe, Nick. The paradox of thrift vs the paradox of hoarding. Worthwile Canadian Initiative, October 01st, 2010. https://worthwhile.typepad.com/worthwhile_canadian_initi/2010/10/the-paradox-of-thrift-vs-the-paradox-of-hoarding.html
3 King, Robert G. 1993. “Will the New Keynesian Macroeconomics Resurrect the IS-LM Model?” Journal of Economic Perspectives, 7 (1): 67-82.DOI: 10.1257/jep.7.1.67
4 Prasad, Pradyumna. Why you should issue GDP linked bonds, Bretton Goods Substack, July 20th, 2021. https://brettongoods.substack.com/p/why-you-should-issue-gdp-linked-bonds
5 Sumner, Scott. Money and output (The musical chairs model), The Money Illusion Blog, April 06th, 2013. https://www.themoneyillusion.com/money-and-output-the-musical-chairs-model/
Subscribe
Notify of
guest
0 Comments
Inline Feedbacks
View all comments