GDP growth rate forecast

[myriam2222]

Dear all,
I decided to resume studying for Level III and I am hoping that being active on the forum will bring me some motivation.
Regarding GDP growth, two ways of breaking it down are presented in the curriculum (maybe it’s merely one and the same approach and that’s actually my question):
Approach #1:
GDP growth breaks down into:
(a) Growth from labour inputs, comprising;

• Growth in potential labour force
• Growth in labor partcipation

(b) Growth from labor productivity, comprising;

• Growth from capital inputs
• TFP growth

Approach #2:
Growth in GDP = %DTFP + alpha * %DK + (1-alpha)*%DL
where alpha = output elasticity of capital
L = labor
K = capital
How would the first approach actually translate into a formula (are we actually supposed to translate it into a formula)?
Are the two approaches different, potentially contradictory? The first one does not take into account output elasticity of capital and labor.
Actually, to be more precise, in the curriculum the first approach applies to GDP trend growth while the second approach applies to total real economic output growth. But I don’t know if this is relevant. Does it make any difference? I mean, I know the difference between trend growth and real GDP growth but why wouldn’t approach #1 also apply to real output growth and approach #2 to growth trend? The only difference I could see between looking at growth trend versus real growth would be the values of the inputs chosen, not the formula itself.
Thanks in advance for your help!

 

[lammy]

myriam2222, your approach #1 & #2 both seek to measure the growth rate in potential GDP.
I typically write the equation for approach #1 as growth rate in potential GDP = long-term growth rate of labor force + long-term growth rate in labor productivity.
The advantage of approach #1 is that you don’t have to estimate capital input or total factor productivity. It incorporates both capital deepening and TFP progress in the labor productivity term.

 

[myriam2222]

I see, thanks.
It remains that both approaches seem slightly different to me, especially as the first approach does not consider output elasticity, which in practice (if my understanding is correct) is the fact that GDP will not grow proportionally to the increase in labor input (respectively capital input) if capital input does not increase proportionally (respectively labor input).
Since they do not give the formula related to the first approach in the curriculum, I guess we will always need to use approach 2 in the case of a calculation, and the first approach may rather be used in a descriptive context.

 

[Curmudgeon]

This may be inaccurate, but here’s my take.
If working for a macro fund, for example, you’d be more likely to use approach #1 to project GDP growth since it involves inputs a firm could reasonably develop projections for.
Approach #2 is more of a theoretical approach since (if I recall correctly) the %DTFP is a plug figure used to describe the component of GDP growth that can’t be explained by capital and labor. As such, approach #2 is less likely to be used for projections in practice, though if given all but one input on the exam, you should still be able to solve for it.
It could be said that Approach #2 is accurate for descriptive purposes (since we know that there is a labor component, a capital component, that there isn’t a 1:1 relationship between growth in labor or capital and growth in GDP, and a technological progress component), but is less useful for projections than approach #1.
Do you agree?

 

[myriam2222]

Hi,
Thanks for the comments.
I have no view on the usage of these approaches by analysts in practice, but I can only say that your reasoning makes perfect sense. And somehow it matches with what Lammy said about the fact that TFP is the problem (the value that we can’t know) and that approach #1 advantage is to avoid having to use it.
So approach #2 would be rather used for backwards solving of %DTPF when we actually already know GDP growth. This is what you mean right? And probably also for backwards solving of alpha because I guess it is just as problematic.
Makes perfect sense, thanks!