Data needed to study macro and financial quantities (random thoughts)

Okay, what does one need to study a market?  It requires many basic components, the first block consists of 5 series:

A. Domestic Production
B. Imports
C. Exports
D. Inventory Levels
E. Price

These basic items are used to derive domestic demand:

Demand/Consumption = Prod + Imp - Exp - Change in Inv

Think meat consumption (everything in lbs).  To derive domestic demand: we add the slaughter of animals; add in the imports and subtract the exports of meat and meat products; and subtract the change in cold-storage stocks. 

think of GDP = C + I + X - M + Delta Inv

Consumption/Demand for Consumer Goods = Production of Consumer Goods + M for Consumption - X for consumption  - Change Inventory (Consumer Goods Inv)

Sales for Consumer Services = dollar volume of services sold,
Service Production = total man hours worked. average dollars per hour
Service Profits = Service Sales - Man hours worked * Wage ($/hr)

What is the proper definition one should use in constructing per capita consumption?

Pop (t)  = Domestic Pop (t-1) + [Domestic Pop (t-1)*birth rate(t) - Domestic Pop (t-1)*death rate (t)] + New immigrants (t) + Emigrants (t) + Tourists - Americans Abroad

One should be as precise as possible when defining population as a per capita divisor for domestic consumption of services and perishable goods.  It also helps in other ways.    Perhaps with a more precise definition of productivity.  For example, retail sales should be divided by the number above.

Now what about Residential Construction?
I would drop tourists and keep the definition in-tact.  Also  I would want to divide on a unit basis not on a dollar volume.  Yet, upon further reflection, "Residential Consumption" is actually what? Change in Inventory? New Home Sales? and New Residential Construction?  And what should be the divisor? To take advantage of cross-sectional evidence, divide new construction by the number of couples + single parents with children.  Take consumption shares and apply that to your population data with demographic detail.

I guess my random thoughts here are about measurement in a way.  In a way, I am concerned about measurement.  Where do the guidelines for measurement come from? Where else, theory and past empirical experience.

We often talk about productivity in hour terms, but should we not also discuss retail sales in hours as well? The dollar volume of retail sales per domestic person per hour.

So we can add to the list:

F. wage rate per hour
G. total employment
H. avg. hours worked
H2. number of unemployed/job openings for the sector

these things give us the ability to look at Productivity and Labor Costs. Lastly, financial considerations.

I. Debt outstanding of the industry
J. Capital Market Issuance of the sector
K. Sector spreads to long-govts'
L.  Market value and book value of equity
M. Construction of the sector
N. Equipment and R&D spending

With these additional factors, we can compute marvelous things like profits to equity. And we can connect the sector together, from the consumption and production, to the labor market productivity and wage share, to profits to equity, interest service/(gross profits - wage costs).

There is one last thing an industry analysis needs: exogenous factors.  Real dollar index for that sector is a good start.  Perhaps a broader equity index covariance.

One needs a lot of data to run a proper sector-level macro-financial analysis.  However, once the analysis is automated, all this collection makes it worth it.  The past becomes the standard for the guess about the future.  Long-data, detailed data, these things are econometrically advantageous.






Comments