Wednesday, August 14, 2013

Thoughts On All Things Economic

Over at the Big Picture, Barry has been doing a lot of thinking on economists.  His discussion has got me thinking about this topic as well.  Most importantly, I've been thinking about how we at the Bonddad Blog fit into the great pantheon of all things economic.  To explain, let me first provide a bit of personal history.

Neither I nor NDD is a "formally" trained economist.  I first learned about the economy from reading the WSJ in college.  I took (I think) about 15-18 hours of econ in college and in a few courses after college.  I also learned basic econ from various Series 7 courses that I took to get my Series securities license.  When I moved over to fixed income, I had to keep up with economic developments on a regular basis to talk with clients.  And I've obviously kept up with the economy in my legal career as this information is very beneficial to my client inter-actions. 

Over the last few years, on advice of Mark Thoma I've wound up reading a fair number of the early economic writers (Smith, Say, Mill) and some of those who I consider transition economists (those that wrote in the early 20th century) such as Fisher and Marshall.  For post Depression analysis I have two copies of my Samuelson macro book from college along with several international economics texts.

Everything that I've read and studied up to this point really falls under the rubric of "political economy" which was the educational designation of writers such as Smith and Say in the early stages of the development of the economics discipline.  What they sought to do in their writing was to explain at a high level how the economy worked by describing the various actors, the work they did, and the various large components of the macro level economic system.  I remember that as I was reading Adam Smith for the first time I realized that he was describing in detail the various parts of the GDP equation in a fairly detailed manner.  The same can be said for Say and Mill.  Marshall's and Irving's work represents the logical bridge between the earlier economists and later writers such as Samuelson and Friedman, as both Marshall and Irving described in more detail things like price mechanisms and the various events that effect them in the short, intermediate and long term.  There is also the importance of several international text books that explain things like the balance of payments and current account deficit.

What has been distinctly lacking in any of these texts is reliance on complex statistical models.  While there are the general graphs of supply and demand, aggregate level supply and demand and the IS/LM model, all of these graphs show general relationships and flow of funds through an economy.  And this is where I personally "draw the line" as it were.  I've read a ton of economists who base their conclusions on complex mathematical models.  And, frankly, I have an incredibly hard time with that for two reasons.

1.) The US economy is over $15 trillion in size and has over 300 million individual participants.  There is simply no way that your model can account for all the variables in something that big.  It's just not even remotely feasible.

2.) You can make anything look good in a model.  Consider the nature of the "rationale expectations theory" and how it relates to the austerity movement.  Here is an explanation of the rational expectations theory that highlights how incredibly out of touch it is with reality:

Imagine the economy is falling around your ears; you don’t know if you’re going to have a job tomorrow, your partner is already unemployed, you really don’t know about the future, but you really worry about the debt, you just lie awake all night worrying about the debt as people do,

So the government credibly signals that it’s going to massively cut government expenditures and what you do using your rational expectations that are built into your head – well you know the true structural form of the equation governing the economy and the value of the co-efficients therein – big assumption, but put that to one side – you calculate your lifetime budget and lifetime expenditures in relation to the fact that twenty years from now that because of these state spending cuts now you’ll pay less taxes then.  Thereby you can retrodect how much extra money you’ve got now and everybody goes to Ikea and buys a couch and that cures the recession.

I am not making this shit up.

That's economist Mark Blythe.

An entire economic argument in favor of austerity was based on the above rationale.  Raise your hand if you think anybody thinks that way.

And most importantly, not only were the analytical underpinnings of austerity proven wrong, the application thereof has been an absolute disaster as demonstrated by both the UK and EUs' experience over the last three years.  And this in spite of the fact the model said it would perform brilliantly. 

A model that tries to show general relationships and general flows of funds is worth looking at.  But complex models that claim to model the entire US economy just aren't worth the time of day no matter how good the algorithms backing it up.

So, let me return to the idea of what we can do and how that applies to our readers.

1.) We can observe general trends in the economy.  We do know that there are leading, coincident and  lagging economic indicators.  We know because there has now been enough history and statistical record keeping to show these relationships.  In watching these trends we can make some generalizations about potential directions the economy will take.

(Completely as an aside, go to the Conference Board's website and look at the different leading and coincident indicators for different economies.  It's really interesting)

2.) We can essentially play the odds and say things like, "with all/most of the leading indicators doing x, we can say with a fairly high degree of certainty that y will happen.  It's not a forgone conclusion, but we can say there's a distinct possibility.

3.) We can show correlation, but in doing so we must understand this may not be causation, or that the causal relationship is not as strong as the data implies.

4.) We can look at the data and the historical record to show you what happened when.  The US -- especially since WWII -- has been keeping a fairly detailed statistical record of the US economy's overall performance.  There are also some pretty good anecdotal records (The Fed's annual summary, the Economic Report to the President) that give us fairly decent context.  

5.) We can make general calculations about the economy.  For example, we know the GDP equation and we also have a fairly decent idea about multipliers.  So we can do things like, "if we increase government spending by x, with this multiplier, we should get y."  We can do the same thing with other GDP components.

6.)  We must understand that we can be wrong.