Never reason from a price change
Many of my students think that we can expect consumers to buy less gasoline when the price is high, and vice versa. In fact, the opposite has more often been true in recent decades, as oil demand shocks have become more important that oil supply shocks. More oil was being consumed when prices hit $147 (due to high demand) than a year later when prices had fallen in half (due to lower demand.)
Paul Krugman’s much too smart to make that mistake, but he comes close in this post:
Tim Duy has an interesting point about the impact of the euro crisis on the United States: in the short run, it may help, not hurt, our prospects for recovery.
Many of us have noticed that the US exports only a bit more than 1 percent of GDP to the euro zone, limiting the direct negative impact. Meanwhile, as Duy points out, the immediate impact of the euro crisis has been (a) a fall in oil prices (b) a fall in long-term interest rates. Both of these are actually positive for the US.
Technically this is correct, as Krugman and Duy are only looking at the partial effect of lower prices on imported oil. The euro crisis has clearly lowered expected NGDP growth in the US and indeed throughout the world. The lower expected NGDP growth will tend to reduce long term interest rates and commodity prices. And it is true that for any given level of AD, lower oil prices are expansionary (as AS shifts right.) But at a deeper level it is very misleading.
There is a reason why both long term rates and commodity prices are strongly correlated with severe AD shocks; both reflect the fact that output of manufactured goods and investment demand are highly cyclical. When NGDP fell sharply below trend in late 2008 and early 2009, both (real) commodity prices and nominal long term rates fell sharply. Those were not hopeful signs; rather they were indicators that AD was coming in far below target.
So yes, it is technically true that, ceteris paribus, lower oil prices and long terms rates can be helpful. But if you wake up in the morning and pick the paper, and see this headline:
Oil Prices Plunge, Treasury Bond Prices Soar.
You should be afraid, very afraid.
My suggestion is that people should never reason from a price change, but always start one step earlier—what caused the price to change. If oil prices fall because Saudi Arabia increases production, then that is bullish news. If oil prices fall because of falling AD in Europe, that might be expansionary for the US. But if oil prices are falling because the euro crisis is increasing the demand for dollars and lowering AD worldwide; confirmed by falls in commodity prices, US equity prices, and TIPS spreads, then that is bearish news.
The markets are telling us that the euro crisis is hurting recovery in the US, not helping. It’s easy to laugh off the highly erratic markets. But before doing so perhaps we should recall that the markets were also tell us that money was far too tight in September 2008. The Fed ignored those warnings. I have a hunch that right now they’d like to redo the September 16, 2008 meeting.
Krugman’s post was in support of an earlier post by Tim Duy. Duy makes a good point when he notes that the Asian crisis of 1997-98 did not slow growth in America. And I’d have to admit that the markets did not cover themselves in glory during that episode. But I still think the Keynesian approach to AD misses something important. Lower long term rates look like an exogenous shock that is expansionary, whereas I believe they mostly reflect lower growth expectations worldwide (confirmed by falling equity prices.) It’s true that US exports to Europe are modest, but if the strong dollar is symptomatic of unintentional tightening of monetary policy, despite low rates, then it may not be good news at all. An increased demand for dollars will look almost invisible from the perspective of the Keynesian C+I+G+NX approach to AD. (I suppose it would be attributed to less “animal spirits.”) But as we saw in late 1937, a scramble for liquidity can be highly deflationary even if the financial system is in good shape. We are certainly not facing a crisis, but this is something the Fed needs to watch carefully. The last time the dollar soared against the euro was July-November 2008. And how’d that work out?
Tags:
18. May 2010 at 09:07
I have a request/suggestion. Perhaps you could have posts that are titled, “Prediction [Date]: [Topic]”, where you make a prediction about a current economic issue that is different from the prevailing wisdom because of your views on money and prices.
I think this would provide three benefits. First, it would hone your own thinking. Second, if/when you tend to be right, even more people might listen to you. Three, I think your audience would be entertained.
On the downside, of course, your predictions may be wrong because of either unlikely outcomes or misapplication of your own framework, thus causing people to wrongly ignore your ideas. But no guts, no glory 🙂
Just a thought.
18. May 2010 at 11:45
In 1997/98, the Asia crisis (fall in Asian AD) induced a drop in oil and commodity prices. Ad growth in the US remained close to the 5% growth path, so no negative effect on the US economy was felt.
18. May 2010 at 11:48
Scott
Favorable responses to your podcast on the GD.
http://valuingeconomics.blogspot.com/2010/05/something-on-great-depression.html
18. May 2010 at 12:10
The dollar is rising against the euro and even some other currencies but there is a difference between now and fall of 2008. Back then gold fell also and the TIPS market indicated an expectation of outright CPI deflation. Neither of those is true right now. How do you interpret these other market indicators?
18. May 2010 at 12:32
‘The last time the dollar soared against the euro was July-November 2008. And how’d that work out?’
Are you sure you are not reasoning from a price change here? The reason for the Euro falling at this point imo is largely from expected central bank monetary expansion, not from dollar contraction.
18. May 2010 at 13:35
Kevin Dick, My motto is “Good economists don’t make predictions, they infer market predictions.” But I see your point. I am arguing the market predicts better than individuals, even expert individuals. I have been on record from the beginning noting that the markets don’t currently predict high inflation. Right now the TIPS markets forecast barely 1%/year inflation over the next 2 years, and under 2%/year over 5 years. Obviously those numbers won’t be exactly right, but I am on record saying the Austrian pessimists are wrong, high inflation is not just around the corner. In general, however, I am more a critic of the predictions of others, than I am a predictor myself. I believe in the EMH, and hence don’t think economists can predict asset price movements.
Marcus, I agree about 1997. And thanks for the link. Kling also seemed to like it. I can’t bear to watch myself on TV.
Joe, I think the deflationary shock is far weaker this time, maybe less than 1/10th as severe. Strong growth in Asia helps gold prices. So I agree that the 2008 comparisons aren’t exact. I don’t want to oversell the current weakness, unless it gets much worse the US won’t have a double dip recession. But it is a tiny bit worrisome.
Doc Merlin, Good point. But I’d also observe that oil prices are even falling a bit in euro terms, so I don’t think fear of European inflation is the big issue. Commodity prices are weak at the global level in trade weighted terms. I think the increase in the demand for liquidity outweighs any QE in Europe (which is debatable in any case.)
18. May 2010 at 13:51
Scott: Yep. One (maybe two, three) picky exception to your “never”.
I run a firm, producing widgets. I ask myself “What would happen if I raised/lowered the price of my widgets? Would my profits increase?”
The government controls rents. “What would happen if the government raised/lowered legal rents?”
I’m an economist checking stability. “What would happen if the price level increased? Would it create an excess supply of goods causing the price level to fall again to its original equilibrium?”
I’m just being picky. These are cases where the ceteris paribus assumption does hold.
Back on topic. Yep. What is happening in Europe is worrying. And I think the USdollar is rising against most currencies, not just the Euro, which suggests the flight to the world’s most moneyish money. Not good news.
18. May 2010 at 14:13
Nick, Yes, I should have said “never reason from an equilibrium market price change”. I think that addresses your two counterexamples.
That’s a good point about the dollar rising against other currencies, which I forgot to mention. And I also should have mentioned that commodity prices are falling in trade-weighted terms, not a signal that inflation in Europe is driving this process, but rather demand for the “world’s most moneyish money.” I like that expression.
18. May 2010 at 16:35
“I have a hunch that right now they’d like to redo the September 16, 2008 meeting.”
I would also like to believe that. What do you think they would like to have done, in hindsight?
18. May 2010 at 17:32
@Nick:
“Back on topic. Yep. What is happening in Europe is worrying. And I think the USdollar is rising against most currencies, not just the Euro, which suggests the flight to the world’s most moneyish money. Not good news.”
Gold is also rising as well as USD. That is symptomatic of this type of crisis.
@Scott
This is a communication issue, partially Austrians have different definitions than Chicago people.
When an austrian says inflation it means something different from when you say it. It means monetary expansion not cpi increases.
This is because productivity increases are price deflationary but not bad. In a world with no productivity increases, ANY per capita monetary expansion means CPI inflation. In a world with productivity increases we only get CPI Inflation when monetary expansion is large enough to outweigh productivity increases.
This means under price targeting any time there is a productivity increase the federal reserve (and financial system) absorbs some of the benefit of this productivity increase through monetary expansion even when it doesn’t affect CPI.
This is another reason why NGDP targeting is /so/ much better than price targeting. It automatically separates AS caused deflation from AD shock caused deflation and adjusts for AD (which is what monetary policy should do!).
19. May 2010 at 03:02
Doc: good point on gold. It’s noteworthy that the prices of all the other “commodities” are falling, while the price of gold is rising. That should be telling us something.
I have a dumb question about what being a reserve currency really means:
On the “moneyness” of the US dollar: Is it true that, in some sense, if I want to convert (say) Canadian dollars into (say) Australian dollars, I first must convert my Canadian into US, and then US into Australian? (Because that’s what seemed to happen to me in Australia once; I got dinged with 2 buy/sell spreads, not one.)
If it is true, then the USD is the medium of exchange between media of exchange, in some sense. It is meta-money.
Can anybody tell me in what exact sense, if any, what I wrote above is true?
19. May 2010 at 04:06
During the Civil War there was a lot of uncertainty around the USD. Breaking up the US pretty much means breaking up the USD. how was this perceived economically? Did it have a lot of impact on the world-economy?
Maybe we can learn some lessons from this for the current situation in Euroland.
19. May 2010 at 06:42
Declan, At a minimum they would have cut rates. But I think they also might have issued a statement with some sort of promise to “catch-up” later if nominal aggregates fell in the near term. The latter statement would have been more meaningful than the rate cut.
Doc Merlin, So now we have 3 proposed definitions of inflation:
1. mainstream economists define in terms of price rises.
2. Austrians use money supply increases
3. I prefer NGDP increases
Nick, I wish I understood gold prices, but I don’t. Even ex post I don’t see what moves the prices. My hunch is that three factors are at work.
1. Rising Asian wealth.
2. Widely dispersed inflation estimates.
3. Euro currency/banking uncertainty.
Japp, What makes that tricky is that there were two forces at work; war, and dismemberment. War alone might weaken a currency, as there are huge fiscal burdens to fighting the war, which the North financed by printing greenbacks. This forced us off the gold standard. And then there is the problem of the South breaking away. Much hunch is that the fiscal burden was the biggest problem, but the split also probably weakened the Northern dollar.
19. May 2010 at 15:25
[…] ambiguity is precisely why Sumner advises to never reason from a price change. The problem is price and quantity are both related to supply and demand factors. Thus, knowing […]
19. May 2010 at 16:27
“At a minimum they would have cut rates. But I think they also might have issued a statement with some sort of promise to “catch-up” later if nominal aggregates fell in the near term.”
I can believe they would have cut rates below 2% in hindsight, since the statement talks about inflation risks balancing weak growth, and we know what has happened to inflation.
But what sign do you see that they would be willing to “catch-up” ie de facto level target if they had a second chance? If that was so, wouldn’t they be doing it now, and wouldn’t they be telling the ECB to do it (since they are always better at giving advice than taking it)?
20. May 2010 at 05:02
Declan, That’s hard to say. You could argue that now “the horse has left the barn” and its too late to close the door. The purpose of promising to catch-up, is to prevent NGDP from falling in the first place. But now it has fallen. I still think they should catch up part way back to the old trend line (not all the way back.) But my argument today is much weaker than if I had made the same argument (with foreknowledge) back in September 2008.
20. May 2010 at 16:45
I’m not knocking your argument at all. I would love to believe they would level target given a second chance. I’m just wondering if you are basing your hunch on anything more than hope. I don’t watch the Fed (or other central banks) anything like as closely as you do, but I haven’t seen any sign they have moved in that direction. All the unconventional moves they have made (both Fed and now ECB) have been to prevent defaults, not to boost AD in general. Please tell me I am wrong!
21. May 2010 at 00:18
[…] that lower energy costs and interest rates were generally positive for the US. Responses came from Scott Sumner, Felix Salmon, Ryan Avent, and Paul […]
21. May 2010 at 05:53
Declan, Over the past 25 years they have certainly tried to keep AD on target, and often done tolerably well. But I am as puzzled as you, it does seem to have dropped off their radar screen.
23. May 2010 at 01:04
[…] that lower energy costs and interest rates were generally positive for the US. Responses came from Scott Sumner, Felix Salmon, Ryan Avent, and Paul […]
23. May 2010 at 02:06
[…] that lower energy costs and interest rates were generally positive for the US. Responses came from Scott Sumner, Felix Salmon, Ryan Avent, and Paul […]
18. March 2011 at 11:50
[…] Sumner teaches us to “never reason from a price change” and start a step […]
23. May 2011 at 11:03
[…] question in general is “Reasoning from Price Changes”, I did this now and then until Sumner brought my mind to it. The idea behind considering many parts of the Walrasian black box (by looking at a range of […]
3. December 2011 at 16:11
[…] aphorisms I’ve come across originates with the very wise and only occasionally obnoxious Scott Sumner: “Never reason from a price change.” Prices don’t just change on a whim. If a […]
15. March 2012 at 04:44
[…] Never reason from a price change. Instead of beginning analysis by thinking about which way prices move, try to figure out whether it was a supply or demand shock and which way the curve shifted. This will avoid logical errors and circular reasoning. […]
14. February 2013 at 09:51
[…] This is what Scott has to say on the issue: My suggestion is that people should never reason from a price change, but always start one step earlier””what caused the price to change. If oil prices fall because Saudi Arabia increases production, then that is bullish news. If oil prices fall because of falling AD in Europe, that might be expansionary for the US. But if oil prices are falling because the euro crisis is increasing the demand for dollars and lowering AD worldwide; confirmed by falls in commodity prices, US equity prices, and TIPS spreads, then that is bearish news. […]
22. March 2013 at 07:37
[…] it may be mistaken for the independent variable. No one thinks that quantities are causal, but some people do mistakenly think that prices cause things. Putting P on the vertical axis is a subtle reminder that no, prices don’t cause things. They […]
17. April 2013 at 04:01
[…] A related mental mistake is to confuse changes in quantity consumed with shifts of a curve. Just because quantity consumed changed, doesn’t mean that it was because the supply curve shifted, nor because the demand curve shifted. A useful caveat is “Never reason from a change in quantity” (the close cousin of the also-very-useful caveat “never reason from a price change“). […]
26. September 2013 at 04:04
[…] rant: Saying that “rising house prices is bad” is reasoning from a price change. Because Japan. Japanese residential real estate prices collapsed by about 50% since 1991. […]
22. December 2013 at 19:16
[…] consequences depends on what is happening (and expected to happen) with money. Kohler is breaking Sumner’s dictum–never reason from a price change. (Or, in this case, a price level change). Price, it is a […]
14. October 2014 at 04:11
[…] the confusion. It is hard to differentiate between supply and demand shocks, but we should never reason from a price change and Scott Sumner is therefore totally correct when he is saying that we need a NGDP futures market […]
9. July 2015 at 09:20
[…] It all seems rather straightforward, but Ip has mistakenly reasoned from a price change. […]
8. September 2015 at 19:51
[…] it may be mistaken for the independent variable. No one thinks that quantities are causal, but some people do mistakenly think that prices are. Putting P on the vertical axis is a subtle reminder that prices don’t cause […]
1. March 2017 at 02:32
[…] then (and only then) does car ‘A’ make more sense as the purchase preference. Never reason from a price change is my advice here, in the case of both cars and stocks; we must consider the change in underlying […]
22. March 2017 at 17:01
[…] There are also some problems with thinking in terms of interest rates more generally. Lowering interest rates could be expansionary, as Rogoff suggests. Or, it could be contractionary—indicating that the monetary authority is committed to keeping inflation low in the future by slowing the growth rate of money. It’s hard to tell which is which. Likewise, falling rates could indicate that the supply of loanable funds has increased—meaning more investment spending. Or, it could indicate that the demand for loanable funds has fallen—meaning less investment spending. For this reason, Scott Sumner quite sensibly advises against reasoning from a price change. […]
13. September 2017 at 02:28
[…] reincarnations, the GSFCI still treats changes in the variables as absolutes. (See Scott Sumner: Never reason from a price change.) If the 10-year Treasury yield were to fall below the funds rate, inverting the yield curve, the […]
13. April 2020 at 09:37
[…] is endogenous to expectations of the size of future income. As with Scott Sumner’s dictum to “Never reason from a price change”, positing an endogenous variable changing autonomously leads us to tell causally confused stories. […]
2. May 2022 at 20:31
[…] https://www.themoneyillusion.com/never-reason-from-a-price-change/ […]
29. June 2023 at 07:15
Dear Scott,
Can you please email me?
Regards,
Ron Baldwin
Brisbane
Australia