Asset markets are economic historians
In a recent post, David Henderson makes this important observation about history:
If we [historians] can’t speculate about what would have happened if x hadn’t happened or if politician y hadn’t been been in charge, then what’s left other than recitation of facts?
I like to think of asset markets as economic historians. When new information about economic fundamentals enters an asset market, then the change in asset prices can be viewed as the market’s verdict on this counterfactual question:
What is the value of a given asset with and without that change in economic fundamentals?
This is exactly the question that historians should be trying to answer. Even better, asset markets don’t just provide one man’s opinion, but rather (if the EMH is true) the optimal view of the relevant counterfactual. Unless economic historians think that they can beat the markets, their only role is to try to decipher the meaning of asset price movements, not second-guess investors.
Of course asset prices aren’t able to answer every question of interest to economic historians, but they are able to answer many of the questions. For instance, both old Keynesian and extreme New Classical economists deny that monetary policy can impact the real economy once interest rates fall to zero in a deep depression. The fact that real asset prices did respond strongly to all sorts of different monetary policy instruments, both before and after the US left the gold standard in 1933, shows that these theories are probably wrong.
I’ve always regarded Friedman and Schwartz’s Monetary History as the best book ever written on macroeconomic history. But it has a huge flaw; they mostly ignored the impact of monetary policy shocks on assets prices. Thus they threw away an enormous amount of useful information on policy counterfactuals. I’m not sure why they did this, but my guess is that relatively early on they noticed that the asset markets did not agree with many aspects of their account of the Great Depression.
So why was the Monetary History so successful despite this methodological flaw? Perhaps because Anna Schwartz is an extremely talented researcher, and because in 1963 Milton Friedman had a deeper understanding of monetary economics than anyone else on the planet.
Tags:
2. May 2010 at 10:31
Well, it was successful because it did make an important, although incomplete and imperfect, contribution (I think there are other problems with their story besides ignoring asset prices also). Seminal contributions are always imperfect, but they capture important aspects of a story and so are highly respected. For example, it was almost immediately known that the Bohr Model was incorrect at its statement of the ground state of orbital angular momentum, and it was quickly known it was insufficient to calculate the emission spectrum of anything but the smallest atoms. That aside though, it was an extremely important step towards the development of a quantum theory of the atom, and eventually to quantum mechanics, and deserves it’s role as an important contribution in scientific history. A story doesn’t have to be perfect in order to gain success and respect, it just has to capture important insights that weren’t fully recognized prior.
On the note of asset prices, I would think it may even be difficult for economic historians to decipher asset price movements in times of crises. I suspect that in financial crises the EMH doesn’t hold up particularly well (I think there is a lot more evidence that the EMH is a long-term phenomena, rather than particular short-term movements). I think in the midst of a financial crisis, two things probably happen. Most investors probably become very uncertain about the conditions of the market and, I suspect though it’s hard to capture evidence for this, that Keynesian “animal spirits” begin to take over and that many rational actors that are required for the EMH to approximately hold begin to act more like noise traders as they revert to their psychological instincts in times of crises (and we now know that it’s not always possibly to arbitrage noise traders mistakes away so the market will deviate from the EMH story). Granted, this is not entirely relevant for it in the context of the Great Depression since these psychological freak outs are likely transitory rather than prolonged events like the Great Depression was, but I think these effects would be relevant during 1929-1931 possibly.
2. May 2010 at 11:05
Ted, I agree with the Bohr analogy.
I have exactly the opposite view regarding the EMH. I think it works best for extremely short term changes in asset prices. In the Great Depression I was able to find lots of examples where asset prices responded immediately to policy news. I have no reason to think the responses were irrational. Over the longer term it is more difficult to disentagle all the various factors that could be influencing asset prices. But if stocks jump sharply in the first 5 minutes after each Fed interest rate change, you can be pretty sure what is moving asset prices in those 5 minute intervals.
2. May 2010 at 11:37
Not necessarily. Instead of giving you some sort of optimal calculation based on probabilities; in general markets should give you a risk-weighted assessment. Bad states are overweighted.
I think the same holds true when you ask people the probability of bad rare events. On a risk-adjusted basis, you should care about terrorism more than by the probability at which it happens.
2. May 2010 at 12:42
“I have exactly the opposite view regarding the EMH. I think it works best for extremely short term changes in asset prices.”
During crises stockmarket works really well in identifying AD shocks. But if you are interested in something else, then you have a problem, as during panics correlations between assets move sharply up, and market processes non AD shock related information less efficiently.
2. May 2010 at 13:24
You got the wrong impression, I think, from what I meant since I wrote the first sentence incorrectly. My first sentence should have been “On the note of asset prices, I would think it may even be difficult for economic historians to decipher the validity of asset prices in times of crisis.” I’m not sure why I wrote “movements.” Asset price movements are somewhat distinct from the efficient market hypothesis. All the EMH says, essentially, is that asset markets are informationally efficient. Whether asset price respond to specific policy movements is more or less irrelevant to whether asset prices are valued “correctly” in the sense of that they reflect all relevant information. My contention was that in a financial crisis the asset prices may not be reflecting all such relevant information.
I certainty don’t deny that asset prices movements are valid and rational, but whether those asset price movements are valid and whether their valuation is valid in an informationally efficient sense are two separate questions. For example, I believe the movement of asset prices in the summer and fall of 2008 were perfectly valid, but I don’t think it’s unreasonable to suggest that the market lost a lot of its informationally efficiency that wasn’t regained for several months (and I think you can have a debate about how efficient the over-the-counter derivative market really is to begin with).
Perhaps you disagree with my correction still, but to be clear I believe asset price movements are not invalid or irrational , but that their valuation may not be based on information efficiency.
2. May 2010 at 15:56
Thorfinn, Yes, but that is part of what you need to take into account when drawing inferenses from asset prices. In practice, risk premiums tend to be pretty small in most asset prices. In any case, what’s generally most informative is changes in asset prices, not levels.
Ted, I think I understand your position better, but I still think asset markets are useful as long as the price changes are informationally valid, to use your term. I don’t care if stocks are 15% overvalued, so long as that is true both before and after an important news item impacts the market.
The information is asset markets was extremely useful in late 2008, unfortunately policymakers choose to ignore that information. Markets signalled that unless money was made much easier, NGDP grwoth would come in far below any plausible target. The Fed ignored the markets and now we are paying the price.
2. May 2010 at 17:23
I guess because you mentioned the EMH I assumed you are talking about the informational-quality of its valuation since you can easily believe that asset prices move in response to policy actions without believing in the EMH at all. Also, certain assets would still be useful – I don’t anticipate the whole market to flip out. Movements would still, of course, be useful for framing policy even if the valuation wasn’t perfect – the trend is still going to be correct (and if you overshoot a bit from it anyway, so what? You averted a recession and you can calm it down rather easily).
As a side note, speaking of movements rather than actual values as being meaningful and the Fed’s performance – you can actually even tell the Fed dropped the ball in June / July 2008 by looking at something so basic as an inflation expectations survey. The actual future inflation projected by consumers is usually pretty imperfect, but it’s actually an excellent indicator of movements in the beliefs of future inflation. That’s why I find it funny at around June / July 2008 inflation expectations just nosedived. Take a look at this graph:
http://research.stlouisfed.org/fred2/graph/fredgraph.png?&chart_type=line&graph_id=25105&category_id=0&recession_bars=On&width=630&height=378&bgcolor=%23B3CDE7&graph_bgcolor=%23FFFFFF&txtcolor=%23000000&ts=8&preserve_ratio=false&fo=ve&id=MICH&transformation=lin&scale=Left&range=Custom&cosd=2007-06-01&coed=2010-03-01&line_color=%230000FF&link_values=&mark_type=NONE&mw=4&line_style=Solid&lw=1&vintage_date=2010-05-02&revision_date=2010-05-02&mma=0&nd=&ost=&oet=&fml=a
Now, I’m not suggesting that we use surveys at the sole guide of monetary policy – but I think that an inflation-targeting Fed should have been at the very least concerned when they see the expectations take a nosedive (ignoring their projected future inflation value – it’s the trend that matters).
Basically, I just find it amusing a survey can even show you the Fed really screwed up.
2. May 2010 at 22:14
“I’ve always regarded Friedman and Schwartz’s Monetary History as the best book ever written on macroeconomic history. But it has a huge flaw; they mostly ignored the impact of monetary policy shocks on assets prices. ”
So true!!!
Elasticity of substitutability of an asset with money for the purpose of “store of value” is important!
This is why monetary expansion makes some asset prices rise more than others at the same time as interest rates are falling. My intuition is that if you combine this with price level rate of change targeting and different productivity increases in different industries you will get these odd effects.
1. The dispersion in rate of change of the price level in the basket of goods used to measure the price index will increase. The goods that are increasing in price will increase more in response to monetary expansion than the goods that are decreasing in price. Also, when the money supply contracts, the distribution of the change in price levels (that make up the basket) will tighten and slide down ward.
2. This means if EMH is true, we will get something that looks a lot like price bubbles in assets that have high substitutability with money for “store of value” and are not denominated in dollars (land, precious metals, etc).
3. This means that using the mean price level of a basket of goods (or the growth rate) as the target of monetary policy is inherently unstable.
4. A temporary solution is NGDP targeting of some sort.
2. May 2010 at 22:15
Oh, I forgot to mention, Hayek talked about monetary policy shocks effect on asset prices.
3. May 2010 at 05:46
123, I’d want to consider the point you raise on a case by case basis. I agree that I am udsually thinking about AD shocks. But I have seen people use asset prices in all sorts of creative ways. For instance, the way other firms in an industry see their share prices respond to mergers of two firms. The respionse of non-merging firms share prices discriminates between the “efficiency” and “monopoly” theories of mergers.
Ted, That’s a good point about the Michigan survey. It is biased (I think it over-forecasts inflation), but does pick up changes. I’m guessing that many people were focusing on the big shifts in gasoline prices during the period you mentioned. They peaked at over $4 in mid-2008, and then fell sharply. I prefer the TIPS spreads. The 5 year TIPS spread had fallen to 1.23% right before the mid-september FOMC meeting. At the meeting the Fed decided not to cut rates, indicating it was still worried inflation would be too HIGH.
Doc Merlin, Yes, that’s a good point about how nominal shocks cause increased price dispersion. You will get bubble-like patterns in commodities, even if the EMH is true.
Doc#2, Yes, the interwar economists (not just Hayek) did a better job of focusing on the asset markets.
3. May 2010 at 09:34
AD shock analysis always works well. Other analysis (mergers etc.) might fail during panics.
3. May 2010 at 13:29
I have a question, since the topic of the EMH came up: can one say a weak form of the EMH states that without external forces (policy, weather, etc) the probability distribution of future (the next in the time series) asset prices is peaked at the current asset price? Or even more weakly, that the probability distribution has no other future asset price at a higher probability than the current price?
Is this an accurate assessment?
3. May 2010 at 13:49
@Scott
I obviously prefer TIPS rates also, I just found the survey interesting. And yes, probably oil price drops are playing a role, but I still think it reflects a general consumer sentiment that inflation wasn’t expected to be common in the future.
4. May 2010 at 06:25
123, Yes, but I didn’t mean merger studies during crises, I just meant that mergers studies are something else that could be done with asset prices. You might want to use 30 years worth of data, in which case you wouldn’t need crisis years.
Jason, That may be approximately true in most cases, but I can’t see how it would be exactly true. The world is a very messy place and doesn’t fit our models exactly. And there are a few cases where it won’t even be approximately true–but finance isn’t my area so I can’t think of one off the top of my head.
Ted, I agree.
6. May 2010 at 11:45
@Jason
“I have a question, since the topic of the EMH came up: can one say a weak form of the EMH states that without external forces (policy, weather, etc) the probability distribution of future (the next in the time series) asset prices is peaked at the current asset price? Or even more weakly, that the probability distribution has no other future asset price at a higher probability than the current price?”
Note quite right. It depends on how much profits you can make from other investments. Weak form EMH says that you “can’t on average beat the market” unless you have insider info or something similar. It doesn’t preclude the entire market moving growing in such a way that makes you money. Being long the DJIA since it was founded, for example, would have made you a lot of money, but it wouldn’t have yielded higher than market returns.