Consumption — 2 ECN 201 -Economic Data Analysis

Consumption - 2 ECN 201 -Economic Data Analysis Lawlor

Review • Linear form of consumption function: – C = a + b*Y – where a is “autonomous spending” – b = MPC – Fits Keynes’s “Fundamental Psychological Law” – Fits the U. S. post-war data: Gretl • LR MPC =. 96 – Fits the U. S. Great Depression • MPC =. 76

Further Explore The MPC APC = the Average Propensity to Consume – = C/Y = (a + by)/Y = a/y + b • Define MPC and APC graphically • Show to simulate it on Excell

Explore meaning of “a” • Statistically is the measure of our ignorance • Algebraically, Graphically a’s arithmetic sign related to the rate of change of APC while income changes – If it is negative, a/YMPC, APC>MPC, and rises as income rises – If it is zero, a/Y = MPC, APC = MPC, and is constant as income rises

Can Simulate this with Excel • Decide on the parameters “a” and “b” • Fill in different values of Y to get different values of C • Show on Excel

Does the l. r. Cons. Fn. Make Sense • Remember in Keynes’s discussion of “units” he said aggregate variables were only comparable over the s. r. – Meant aggregate Y, C, P • We may want to restrict our comparisons to “decade” long units at the most – We ask you to do some of this type of modeling in Gretl in your assignment - show

Or, both in the s. r. and the l. r. may want to add more variables • Keynes mentioned “windfall changes in capital values” • Perhaps the way Americans, and American society, has altered the way that average people save, and so changed the average household balance sheet, has effected the MPS • Discuss this, and the difference between “defined benefit” plans and “defined contribution plans”

What sorts of assets are important to Households • Mutual Fund saving suggest that financial market instruments might be important – Could proxy this effect statistically with stocks • See the St. Louis Fed database • The Current period suggest some people are saving through speculation in housing

Note relation to cons. Fn. • We are saying there is more than one variable responsible for a con. fn, and so a savings fn. – C = f(Y, stock values, house values, culture, financial market innovation (sub-prime mortgages)) – Statistically, this increase in rhs variables means we are entering into the realm of “multiple variable regression”

One linear form of this • C = a + b 1 Y + b 2 S&P 500 • Note we are asking you to explore this possibility in the LR in your project • Look at TRSP in Gretl, graph it as a time series, explain its dynamics • Explore time series plot alone and with C, with APC • Show to construct APC • Class example might be to consider a recent s. r. period: – Note: first consider data availability, what we know of institutional and economic history

• TRSP Limits us to 2003: 12, so lets run a regression for the period 1983: 12 – 1993: 12 – Captures the period of run up in stock values • Show in Gretl “range” and “model” • Model the cons. fn. for this period with and without TRSP