The test version of the NGDP expectations system is running. A graph of the forecast in y/y % change format can been seen at this link.
Let me know in the comments if you’d like the data shown differently and I’ll see what I can do.
Big thanks to Lars “The Market Monetarist” Christensen for the shout out.
In this year of the snake, output of goods and services has been weakish
but hiring has been strongish
because payrolls are a medium term proposition, and NGDP expectations are fairly upbeat (for this demand-starved recovery)
For context, note that nominal GDP (the quantity of spending in the economy, which the Fed determines by its money printing decisions and signals thereof) has grown around 3.5% per year. 4.5% nominal growth is about as strong as we’ve seen in 6 years.
Firms are hiring in anticipation of moderately stronger output in 2014. It takes six months before a new employee adds value, a year before they hit their stride. Who cares if sales growth slows today? As long as the light bill is paid, firms look to tomorrow. If the private sector views the sequestration as a temporary drag, they’ll go ahead and hire as though the sequestration never happened. The Fed wears the pants.
I just read the markets and the Market Monetarist model fits reality better than anything else out there.
P.S. output wasn’t even that weak in 2013H1. If you go by real GDI (which uses the same price deflator as real GDP), growth was typical for the recovery, a bit north of 2% y/y.
Lars E.O Svensson.
Why not? If we can have a creepy outlander like like Henry Kissinger hold a high position, why not a charmer and super qualified foreigner like Lars Svensson? Someone who understands that debt-to-income ratios only scratch the surface and who we could trust to target the forecast. Someone who could bring some practical financial regulation to the U.S. too. What better way to spin “immigration reform” than with the example of an immigrant who brought monetary stability to America?
I think Mankiw would be a great choice, but he technically worked for Bush and so stands no hope of being nominated by Obama, thus Svensson is my next pick.
Svensson is still a long shot but his background makes me think he could make it through the confirmation process if only given the chance. First, he is not American. This should help pull in Democratic votes. Next, he is from a nonthreatening white ethnic group, this should help pull in Republican votes, because as the New York Times teaches, all Republicans are mouth-foaming racists.
I doubt Svensson is on anyone with power’s list. Still, I hope we can get a real economist at the Fed, someone with New Keynesian credentials. New Keynesians can at least talk to Market Monetarists, more than can be said for Vulgar Keynesians like Summers. We’d do well to keep in mind that, so long as the Euro is kicking around, the U.S. money supply is at risk of being snatched up. This means we need someone with Swiss resolve to accommodate this prospective demand with freshly printed green paper. I do buy the argument that the median macro economist mostly sets the policy space for the Fed, but its hard to imagine Larry Summers getting the response to a hypothetical Euro Zone collapse/scare right. The thought of Summers steering the ship makes me shudder.
Ben Bernanke was amongst the most respected macro researchers when he took the throne in 2006, a peer of Svensson, Gali or Mankiw. Larry Summers on the other hand has rather less impressive monetary credentials. He’s of course a smart man, but in all the wrong ways to be Fed chairman. It doesn’t matter how high your IQ is, if you get money wrong, you can’t get much right in macro. Frankly, I see Summers as just another high priest in the progressive cult, at least in the monetary policy field. I respect him for understanding Gaussian statistics, but he shouldn’t be chairman.
As flawed as Bernanke has been, it will be hard to replace him with the names floating around in the press. Bernanke owes it to humanity to stick around through the end of Obama’s term in case a more moderate Democrat can be elected in 2016, and someone more Mankiw-like put on the table of potential replacements. (or Chris Christie can discover ketogenic diets, win, and pick Mankiw).
I have this nutty, gnawing idea that there are fewer bad deals in this world than most think.
I came to this position by borrowing from Robert Shiller and a few others. Actually, it’s possible this very argument was posted on some well known econblog and I’ve unknowingly stolen it after the seed grew into a mini essay, if so…I’m sure my victim won’t lose sleep.
What I have in mind pertains to markets where one can choose between buying or renting a durable consumption good.
I say that the choice between renting and buying makes no difference, both will generally put the typical consumer in the same position after the sale. Take a random sample of people, make half buy houses and half lease houses, the average net worth of the two groups would be similar at any given future point in the treatment cycle. I think this follows for cars and perhaps other durables as well. I also contend that the two groups will show similar levels of happiness as measured by average heart rate variability, cortisol to testosterone ratios and inflammation markers.
Whenever I try to explain this idea to those uncomfortable with the counterintuitive results which bedevil economics and engineering (the type of results I live for), I’m met with halfwhited responses like “but you don’t GET anything when your lease is up !!!1 LOLOMG[sic]“. This is an excellent example of the old seen and the unseen error. I might rattle off facts such as:
1. Average house prices clearly can fall spectacularly in nominal terms, you’re basically betting on NGDP growth with your lender (or expected resale value with your automaker). Even in good times real appreciation isn’t a given, and you still have to live somewhere. Make a killing on a condo in Manhattan, sell and live like a king in Wichita, you’re still living in Wichita. Note that this is different from buying real estate investments.
2. Local house prices (say in a given neighborhood) pose a big risk. No one can be sure where the government will put subsidized housing (come on it is a real risk to the value) or where a natural disaster will happen (insurance is not always are sure thing). Regular development is a risk too, pay a premium for a grand old farm house near a suburban area, then development overburdens the roads and pulls down your premium, this doesn’t show up in aggregate price indices but still threatens the individual home owner in some cases. You might also have bought a house in say Rome in 400AD, Lindisfarne in 790 or Detroit in 1990, civilizational collapse is never far off. These are risks which must be born or insured against.
3. Houses need upkeep. Not owning a house dissuades one from taking on weekend projects. No one ever said on his deathbed “I wish I’d laid more tile”. Landlords unclog sinks and mow lawns. The tenant pays for this, but then so does the homeowner in one way or another. There is no free lunch.
No free lunch is what my argument is all about.
Here’s what I have in mind for the car market.
Lets say that Ford offers a lease where is the present value of the 36 month lease is the present value of the ‘actuarially fair’ lease and is the markup. I think we can all agree that if Ford charged $20 a month to lease a new sedan, that this would be a good deal. It thus follows that somewhere in money space, betwixt $20 and $1,000 per month, the lease stops being a good deal. This price would be that which leaves the lessor in exactly the same position if they bought a new car on credit and sold it after 36 months or leased the car, drove it the same amount as in the first case, and returned it after 36 months. That is, .
In a seemingly competitive market such as that for autos, I suspect that will tend toward zero. I can’t easily test this, but I can tell you that according to the BLS’ “Urban Consumer” CPI data, new car prices in America have risen 5.7% since 2003, whilst the total CPI has risen 27%. Its hard to tell a story where this happens in a market characterized by monopolistic collusion. Sure, will be slightly different for everyone, because we all have our own personal reckoning on the degree to which we dread dealing with random strangers who might buy our car when we’re done with it. But most will end up in about the same spot, buy or lease. Because people seem so systematically biased in favor of buying over leasing, if anything a lease is more likely to be the good deal. I’m more sure of this for housing than autos.
In the end comes down to lifestyle preferences, as they steer . For myself, I want to be able to pick up and leave within a few weeks if the right opportunity arises. A leased car yields me safe and pleasant transportation, while containing costs if I have to abandon the vehicle in a hurry. Likewise, a month-to-month rental contract spares me the utter drudgery of repairs and yardwork on top of not tying me down with a dear and illiquid durable. But there are probably lots of good reasons to buy as well. Point is, there is no bad deal.
1.Including the nonmoney costs of selling the car in an imperfect market
2. Government policy can also hold sway. My Swedish readers will know well how outlawing market signals can throw a spanner in the gears of a market.
From only one model. The next step is to average over several models (easy). And then to find a way to link this graph, which is updated ever three minutes during the trading day, to the blog (probably easy).
After that things get more complicated. What Silicon Valley friends are for!
Enough already with the monetary policy right? Print money. Keep NGDP growing at a predictable rate. Look at the markets and tell everyone you’re going to. There, another monetary policy blog need never be written.
While we wait for central banks to stabilize the world economy and for politicians to undo that stability with wrongheaded policies, a handful of geniuses are making the world worth living in. These folks are only reason I get out of bed in the morning.
Check out this news clip:
I’ve had a reasonably excessive mancrush on Elon Musk since reading about Tesla Motors back in…2006? Since that time he’s blown past all expectations; replacing NASA, making the 2012 car of the year, founding a solar energy firm, and none of these daring firms have gone bankrupt yet! For the past year he has lead the thinking public on with hints of a an electric airplane and the Hyperloop. Unlike people like me, who have lots of good ideas but are too lazy and dull to carry through on the execution, Musk has already proven himself in a big way. Four times.
I look forward to hearing the details.
BTW. I’ve been chipping away at Nikola Tesla’s autobiography in the last week. Highly recommended, if you’re interested in getting a glimpse at how the mind of this monumental genius worked.
I’m home for lunch, and today has been a rather interesting day in my apartment while I was away.
My computer has been diligently mining financial websites for seven market prices and combining them with data from the U.S. government to generate a new NGDP forecast every three minutes. Technically its an average of NGDP and NGDI.
The procedure is just a prototype, and I haven’t built the infrastructure to show the forecast live on a web page, but it seems to be working. Here is a graph of what I have as of lunch time. The scale is “expected NGDP growth 2013Q3 to 2014Q3. I used autosales, retail sales, payroll employment and personal income to estimate Q2 NGDP. A VECM does the rest.
Not the best graph I’ve made. Sorry for not having the time’s labeled, I’ve made a programming error and my procedure is not recording the times correctly. Still, this is encouraging. The procedure can only get better from here.
Much more to come.
Update: Here is a graph from today, July 9th. The program crashed around noon but I was able to get it going again an hour or so later.
I’d like to alert my readers to a fantastic new data site, Quandl.
The things I like about Quandl are that it:
1. Has loads of macroeconomic and financial time series
2. Has an API, and you can download series directly within the R or Python languages (and a few others).
3.It is free.
BTW, I have no connection to Quandl.
Rather than revise my approach further, I’m going to post a working version of how I propose to estimate market NGDP expectations.
Here is what I have in mind
Take a vector
Representing the price of a set of assets. I propose that this set of assets is strongly correlated with the market’s implicit outlook on U.S. nominal GDP. This is not necessarily a perfect data set, we can imagine that events in East Asia or Europe could steer any single one of these prices for an extended period. Ideally, we’d also have an inflation indexed bond spread, though in this case, this would hinder the analysis as TIPS have only traded since 2003. We might also include a safe private sector bond index.
Take the first and second principal components of for t spanning 1989 to 2012, quarterly frequency. These components might contain much of the market’s near-term outlook, distilled into an index of meaningless scale.
I’ve not labeled the graphs, the black line is the first PC, the red line is the second.
Next, model these components along with lagged nominal GDP in the vector ,using a selection of potential VAR models. We might not be able to say, with any credibility, what the components actually mean, but we can try to forecast nominal GDP with them.
Call these models: . In this example I look at a ‘vector error correction model’ on three lags of using only the first principal component, four lags using only the first, and three lags using both components.
Fit these models to data sequentially, covering period 1 (1989Q1) to as little as 40 (1998Q1) or as much as 97 (2012Q4).
At first, fit model to date index positions 1 to 40, forecast NGDP with this model for period 41 through 45 (remembering that NGDP is lagged one period relative to the market components). Save the percent change in NGDP for periods t+1 to t+5, in the 57×1 vector
Repeat this process over the dataset, in essence simulating the result a forecaster would have had in each quarter of the data set after position 40 (the amount of data needed to give the model a moderately good fit).
Here are the resulting s from three s
That is run off quarterly averages. This looks about right to me, as far as the deviations from 5% approximating the effective stance of policy. Here is the system run at quasi daily frequency, the choppiness of NGDP data make this cumbersome.
You could think of lots of different ways to make the models somehow account for the fact that the principal components are probably distorted by international factors, or otherwise not reliably related to NGDP expectations.
I suggest setting up some sort of Bayesian model averaging or combination system, on a wide set of potential models. That, would be an undertaking.
On and BTW: Link to the quarterly NGDP expectations here.