Saturday, May 28, 2016

No Hesitations at 500K

Some company just emailed to inform me that No Hesitations had made its list of the Top 100 Economics Blogs.  I was pretty happy until I decided that there were probably only 70 or 80 economics blogs.

But seriously, thanks a lot for your wonderful support.  No Hesitations has about 500,000 pageviews since launching in summer 2013, and the trend (below) looks good.  The time has flown by, and I look forward to continuing.

Graph of Blogger page views

Monday, May 23, 2016

Listening to Your Sentences, II

Here's a continuation of this recent post (for students) on listening to writing.

OK, you say, Martin Amis interviews are entertaining, but Martin Amis is not a mere mortal, so what's the practical writing advice for the rest of us? Read this, from Gary Provost (evidently the highlighting is keyed to different sentence lengths):

Sunday, May 22, 2016

Martin Amis on How to Write a Great Sentence

It's been a while since I did a piece on good writing, for students.   In an old post I said "Listen to your words; push your prose toward poetry."  That's perhaps a bit much -- you don't need to write poetry, but you do need to listen to your writing. 

On the listening theme, check out this Martin Amis clip, even if I don't see why you shouldn't repeat prefixes or suffixes in the same sentence (in fact I think the repetition can sometimes be poetic, a sort of alliteration, when done tastefully).  And while you're at it, take a look at this marvelous older clip too.

Friday, May 20, 2016

Hazard Functions for U.S. Expansions

Glenn Rudebusch has a very nice 2016 FRBSF Letter, "Will the Economic Recovery Die of Old Age?".  He draws on perspective and results from our joint work of 25 years ago (including a paper we did with Dan Sichel -- see below), and he applies them to the present expansion.  He correctly emphasizes that U.S. expansion hazard functions are basically flat, so "old" expansions are no more likely to end than "young" ones. That's of some comfort, since the present expansion, which started in mid-2009, is getting long in the tooth!

Actually, the flat expansion hazard is only for post-WWII expansions; the prewar expansion hazard is sharply increasing. Here's how they compare (copied from Glenn's FRBSFLetter):

Probability of an Expansion ending within a month
Probability of a recovery ending within a month

Perhaps the massive difference is due to "good policy", that is, post-war policy success in "keeping expansions alive".  Or perhaps it's just "good luck" -- but it's so big and systematic that luck alone seems an unlikely explanation.

For more on all this, and to see the equally-fascinating and very different results for recession hazards, see Diebold, Rudebusch and Sichel (1992), which I consider to be the best statement of our work in the area.

[Footnote:  I wrote this post about three days ago, intending to release it next week. I just learned that The Economist (May 21st issue) also reports on the Rudebusch FRBSF Letter (see, so I'm releasing it early.  Interesting that both The Economist and I are not only slow -- Glenn sent me his Letter in February, when it was published! -- but also identically slow.]

R/Finance 2016: Applied Finance with R

At R/Finance 2016: Applied Finance with R.  Interesting group, with many constituencies, and interesting program, which appears below (or go to
Friday, May 20th, 2016
08:00 - 09:00Optional Pre-Conference Tutorials
Ross Bennett: Feasible Space Analysis and Hierarchical Optimization with PortfolioAnalytics
Dirk Eddelbuettel: Introduction to Rcpp and RcppArmadillo
Doug Service: Leveraging Azure Compute from R
T. Harte + M. Weylandt: Modern Bayesian Tools for Time Series Analysis
09:00 - 09:30Registration (2nd floor Inner Circle) & Continental Breakfast (3rd floor by Sponsor Tables)
Transition between seminars
09:30 - 09:35Kickoff
09:35 - 09:40Sponsor Introduction
09:40 - 10:20Rishi Narang: Rage Against the Machine Learning
10:20 - 10:50Robert McDonald: The derivmkts package
Piotr Orłowski: Modeling Divergence Swap Rates
Jerzy Pawlowski: Exploring Higher Order Risk Premia Using High Frequency Data
Majeed Simaan: The Implicit Value of Tracking the Market
Kris Boudt: Block rearranging elements within matrix columns to minimize the variability of the row sums
10:50 - 11:20Break
11:20 - 11:40Brian Boonstra: Calibrating Parsimonious Models Of Equity-Linked Default Intensity
11:40 - 12:00Matthew Ginley: Simulation of Leveraged ETF Volatility Using Nonparametric Density Estimation
12:00 - 12:20Klaus Spanderen: Calibration of the Heston Local Stochastic Volatility Model
12:20 - 13:25Lunch
13:25 - 14:05Tarek Eldin: Random Pricing Errors and Systematic Returns: The Flaw in Fundamental Prices
14:05 - 14:25Sanjiv Das: An Index-Based Measure of Liquidity
14:25 - 14:45Ryan Hafen: Interactively Exploring Financial Trades in R
14:45 - 15:03Nidhi Aggarwal: The causal impact of algorithmic trading on market quality
Chirag Anand: Liquidity provision in a high-frequency environment
Maria Belianina: OneTick and R
Patrick Howerter: Connecting QAI to R
15:09 - 15:40Break
15:40 - 16:00Marc Wildi: Monitoring the US Economy: a System of Timely (Real-Time Daily Mixed-Frequency) Indicators
16:00 - 16:06Sile Li: Constructing US Employment Google Search Index by Applying Principal Component Analysis
16:06 - 16:12Doug Martin: Information Ratio Maximizing Fundamental Factor Models
16:12 - 16:18Robert Franolic: Eyes on FX
16:18 - 16:24Warren Durrett: Comparing Private Equity Managers Using an Objective, Data-Driven Approach
16:24 - 17:04Frank Diebold: Estimating Global Bank Network Connectedness
17:04 - 17:10Information about reception and dinner
17:10 - 19:10Conference Reception
19:10 - 19:30(Optional) Transfer to Conference Dinner
19:30 - (Optional) Conference Dinner (Riverside Room and Gallery at Trump Hotel)
Saturday, May 21st, 2016
08:00 - 09:00Coffee/ Breakfast
09:00 - 09:05Kickoff
09:05 - 09:35Hsiu-lang Chen: Do Mutual Funds Exploit Information from Option Prices for Equity Investment?
Kyle Balkissoon: A Practitioners analysis of the overnight effect
Mark Bennett: Measuring Income Statement Sharpe Ratios using R
Mark Bennett: Implementation of Value Strategies using R
Colin Swaney: Evaluating Fund Manager Skill: A Mixture Model Approach
09:35 - 09:55Bernhard Pfaff: Portfolio Selection with Multiple Criteria Objectives
09:55 - 10:15Douglas Service: Quantitative Analysis of Dual Moving Average Indicators in Automated Trading Systems
10:15 - 10:45Marjan Wauters: Smart beta and portfolio insurance: A happy marriage?
Michael Kapler: Tax Aware Backtest Framework
Miller Zijie Zhu: Backtest Graphics
Laura Vana: Portfolio Optimization Modeling
Ilya Kipnis: Hypothesis Driven Development: An Understandable Example
10:45 - 11:05Break
11:05 - 11:25Mark Seligman: Controlling for Monotonicity in Random Forest Regressors
11:25 - 11:45Michael Kane: glmnetlib: A Low-level Library for Regularized Regression
11:45 - 12:05Xiao Qiao: A Practitioner's Defense of Return Predictability
12:05 - 13:05Lunch
13:05 - 13:45Patrick Burns: Some Linguistics of Quantitative Finance
13:45 - 14:05Eran Raviv: Forecast combinations in R using the ForecastCombinations package
14:05 - 14:35Kjell Konis: Comparing Fitted Factor Models with the fit.models Package
Steven Pav: Madness: a package for Multivariate Automatic Differentiation
Paul Teetor: Are You Trading Mean Reversion or Oscillation?
Pedro Alexander: Portfolio Selection with Support Vector Regression
Matthew Dixon: Seasonally-Adjusted Value-at-Risk
14:35 - 15:05Break
15:05 - 15:25Bryan Lewis: R in Practice
15:25 - 15:45Matt Dziubinski: Getting the most out of Rcpp: High-Performance C++ in Practice
15:45 - 16:09Mario Annau: h5 - An Object Oriented Interface to HDF5
Robert Krzyzanowski: Syberia: A development framework for R
Dirk Eddelbuettel: Rblapi Revisited: One Year Later
Matt Brigida: Community Finance Teaching Resources with R/Shiny
16:09 - 16:29Jason Foster: Multi-Asset Principal Component Regression using RcppParallel
16:29 - 16:49Qiang Kou: Deep learning in R using MxNet
16:49 - 17:04Prizes and Feedback
17:04 - 17:09Conclusion
17:09 - 17:19Transition to Jak's
17:19 - Post-conference Drinks at Jak's Tap

Tuesday, May 17, 2016

Statistical Machine Learning Circa 1989

I've always been a massive fan of statisticians whose work is rigorous yet practical, with emphasis on modeling. People like Box, Cox, Hastie, and Tibshirani obviously come to mind.  So too, of course, do Leo Brieman and Jerry Friedman.  

I had the good luck to stumble into a week-long intensive lecture series with Jerry Friedman in 1989, a sort of summer school for twenty-something assistant professors and the like.  At the time I was a young economist in DC at the Federal Reserve Board, and the lectures were just down the street at GW.

I thought I would attend to learn some non-parametrics, and I definitely did learn some non-parametrics.  But far more than that, Jerry opened my eyes to what would be unfolding for the next half-century -- flexible, algorithmic, high-dimensional methods -- the statistics of "Big Data" and "machine learning".  

I just found the binder containing his lecture notes.  The contents appear below.  Read the opening overview, "Modern Statistics and the Computer Revolution".  Amazingly prescient.  Remember, this was 1989!

[Side note:  There I also had the pleasure of first meeting Bob Stine, who has now been my esteemed Penn Statistics colleague for more than 25 years.]

Wednesday, May 11, 2016

Great Yield Curve Graphic

I'm giving an overview lecture today on certain aspects of yield curves and their modeling, which reminds me of this phenomenal NYT interactive graphic.  CLICK HERE to get going, and give it time to load.  Then click "next" to go through nine fascinating graphics, ending with Germany and Japan.  You can also grab and rotate each graphic with your mouse.

Monday, May 9, 2016

On the Origin of "Forecasts"

The word forecasts, that is. From the BBC Magazine,
One hundred and fifty years ago Admiral Robert FitzRoy, the celebrated sailor and founder of the Met Office, took his own life. One newspaper reported the news of his death as a "sudden and shocking catastrophe". Today FitzRoy is chiefly remembered as Charles Darwin's taciturn captain on HMS Beagle, during the famous circumnavigation in the 1830s. But in his lifetime FitzRoy found celebrity not from his time at sea but from his pioneering daily weather predictions, which he called by a new name of his own invention - "forecasts".

And quite apart from the origin of the term, the description of the early development of weather forecasting is fascinating.

[Thanks to Glenn Rudebusch for bringing this to my attention.]

Sunday, May 8, 2016

Safe Assets

Gary Gorton has a fascinating new paper, "History and Economics of Safe Assets", which contains the quote of the week: "...almost all of human history can be written as the search for and the production of different forms of safe assets".  Not sure that's the first cut I'd take at "all of human history", but it's certainly an interesting perspective.  As the old maxim says: "When you have a hammer, everything looks like a nail".

Thursday, May 5, 2016

Unsung Hero: NBER Conference on Research in Income and Wealth

Here's to the the NBER's ongoing Conference on Research in Income and Wealth (CRIW), unsung hero, home of down-and-dirty measurement mavens since 1935.  Yes, since 1935!  Check out Chuck Holten's fascinating CRIW description in the NBER Reporter, and the full list of associated CRIW volumes published. What a stunning record of steady service.

FYI a typical program (in this case, from last summer) appears below.

SI 2015 NBER/CRIW Workshop 
Susanto Basu, Nicholas Bloom, Carol Corrado and Charles R. Hulten, Organizers 
July 13-14, 2015 

Charles B Room
Royal Sonesta Hotel
40 Edwin H. Land Blvd.
Cambridge, MA


Monday, July 13:
8:30 am
Continental breakfast
9:00 am
Thomas Piketty, Paris School of Economics
Emmanuel Saez, University of California at Berkeley and NBER
Gabriel Zucman, London School of Economics
Distributional National Accounts: Methods and Estimates for the United States Since 1913

John Sabelhaus, Federal Reserve Board
10:00 am
10:30 am
Jae Song, Social Security Administration
David J. Price, Stanford University
Fatih Guvenen, University of Minnesota and NBER
Nicholas Bloom, Stanford University and NBER
Till M. von Wachter, University of California at Los Angeles and NBER
Firming Up Inequality

Johannes Schmieder, Boston University
11:15 am
Phillipe Aghion, Harvard University and NBERUfuk Akcigit, University of Pennsylvania and NBER
Antonin BergeaudBanque de France
Richard Blundell, University College London
David Hemous, INSEAD
Innovation and Top Income Inequality

Discussant: Scott Stern, Massachusetts Institute of Technology and NBER
12:00 pm
1:00 pm
Wolfgang Keller, University of Colorado and NBER
Hale Utar, Bielefeld University
International Trade and Job Polarization: Evidence at the Worker Level

David Autor, MIT and NBER
1:45 pm
Dongya Koh, University of Arkansas
Raul Santaeulalia-Llopis, Washington University in St. Louis
Yu Zheng, City University of Hong Kong
Labor Share Decline and the Capitalization of Intellectual Property Products

Discussant: Dan Sichel, Wellesley College and NBER
2:30 pm
3:00 pm
3:30 pm
Muge Adalet McGowan, Organisation de Coopération et de Développement Économiques(OCDE)
Dan R. Andrews, Organisation de Coopération et de Développement Économiques(OCDE)
Labour Market Mismatch and Labour Productivity: Evidence from PIAAC Data
4:00 pm

Tuesday, July 14:
8:30 am
Continental breakfast
9:00 am
Carol Corrado, The Conference Board
Jonathan Haskel, Imperial College London
Cecilia Jona-Lasinio, LUISS University of Rome
Bilal Nasim, Institute of Education
Is International R&D Tax Competition a Zero-sum Game? Evidence from the EU 
Discussant: Bronwyn Hall, University of California, Berkeley and NBER
9:45 am
Antonio Falato, Federal Reserve Board
Jae Sim, Federal Reserve Board
Why Do Innovative Firms Hold So Much Cash? Evidence from Changes in State R&D Tax Credits

Discussant: Daniel Wilson, Federal Reserve Bank of San Francisco
10:30 am
11:00 am
Neil Thompson, Massachusetts Institute of Technology
Moore’s Law goes Multicore: The Economic Consequences of a Fundamental Change in how Computers work

Discussant: Chris Forman, Georgia Institute of Technology
11:45 am
John Bai, University of Southern California
Daniel Carvalho, University of Southern California
Gordon Phillips, University of Southern California
The Impact of Bank Credit on Labor Reallocation and Aggregate Industry Productivity

Discussant: Javier Miranda, Census Bureau

12:30 pm
Joint Session with Macro Productivity:
1:30 pm
Gita Gopinath, Harvard University and NBER
Sebnem Kalemli-Ozcan, University of Maryland and NBER
Loukas Karabarbounis, University of Chicago and NBER
Carolina Villegas-Sanchez, ESADE-Universitat Ramon Llull
Capital Allocation and Productivity in South Europe

Discussant: Diego Restuccia, University of Toronto

2:15 pm
Colin J. Hottman, Columbia University
Stephen J. Redding, Princeton University and NBER
David Weinstein, Columbia University and NBER
What is ’Firm Heterogeneity’ in Trade Models? The Role of Quality, Scope, Markups, and Cost

Discussant: Daniel Xu, Duke University and NBER

3:00 pm

3:15 pm
Lucia Foster, Bureau of the Census
Cheryl Grim, Bureau of the Census
John C. Haltiwanger, University of Maryland and NBER
Zoltan Wolf, Center for Economic Studies, US Bureau of Census
Macro and Micro Dynamics of Productivity: Is the Devil in the Details?

Discussant: Jan De Loecker, Princeton University and NBER

4:00 pm

5:15 pm
Reception at the Royal Sonesta Hotel

Sunday, May 1, 2016

On Forecasting Variation and Covariation

One hallmark of a great idea is that it's "obvious" (ex post). Fantastic recent work by Bollerslev, Patton, and Quaedvlieg (BPQ) certainly passes that test.

BPQ build on the classic Barndorff-Nielsen and Shephard result that the precision with which realized variation and covariation are estimated is time-varying but can be estimated (let's just speak of "variation" for short, whether univariate or multivariate). Put differently, the measurement error in realized variation is heteroskedastic but can be estimated. Hence, for optimal variation prediction, one should presumably weight the recent past differently depending on the estimated size of the measurement error. BPQ do it and get large predictive gains. Check it out here. (This is the new and multivariate (covariance) paper, which cites the earlier univariate (variance) paper.)

Why didn't I think of that? I mean, really, the Barndorff-Nielsen and Shephard result is more than a decade old, and I know it well. Can I not put two and two together? Damn.

But seriously, congratulations to BPQ.