UPDATE: Mea culpa. One point in the original post was dead wrong. It is possible, contrary to what I wrote below, to get something like a 0.7% difference in annual growth rates with the assumptions he has in the chart below (Drum still exaggerated when he called it 1%). I don't know if the model is valid (I have little faith in any macro models) but I was wrong on this claim. Using the 0.7% and working more carefully by quarter we get a cumulative GDP addition a bit lower than the cumulative debt addition. There is still obviously a reasonable question even at a multiplier near 1 whether $1 of economic activity today is worth $1 of debt repayment plus interest in the future.
I am not a believer, obviously, in cyclical tweaking of the economy by the Feds. To my thinking, the last recession was caused by a massive government-driven mis-allocation of capital so further heavy-handed government allocation of capital seems like a poor solution. But what really drives me crazy is that most folks on the Left will seductively argue that now is not the time to reduce debt levels, implying sometime in the future when the economy is better will be the appropriate time. But when, in any expansion, have you heard anyone on the Left say, "hey, its time to reduce spending and cut debt because we need the fiscal flexibility next time the economy goes wrong."
I will leave the stuff in error below in the post because I don't think it is right to disappear mistakes. For transparency, my spreadsheet reconstruction both confirming the 0.7% and with the updated numbers below is here: reconstruction.xls.
I see that Macroecomic Advisors has produced a comprehensive estimate of the total effect of bad fiscal policies. Their conclusion: austerity policies since the start of 2011 have cut GDP growth by about 1 percentage point per year.
Something seemed odd to me -- when I opened up the linked study, it said the "lost" government discretionary spending is about 2% of GDP. Is Drum really arguing that we should be spending 2% of GDP to increase GDP by 1%?
Of course, the math does not work quite this way given compounding and such, but it did cause me to check things out. The first thing I learned is that Drum partook of some creative rounding. The study actually said reductions in discretionary spending as a percent of GDP reduced GDP growth rates since the beginning of 2011 by 0.7% a year, not 1% (the study does mention a 1% number but this includes other effects as well).
But it is weirder than that, because here is the chart in the study that is supposed to support the 0.7% number:
Note that in the quarterly data, only 2 quarters appear to show a 0.7% difference and all the others are less.
I understand that compounding can do weird things, but how can the string of numbers represented by the green bars net to 0.7%? What it looks like they did is just read off the last bar, which would be appropriate if they were doing some sort of cumulative model, but that is not how the chart is built. If we interpolate actual values and are relatively careful about getting the compounding right, the difference is actually about 0.45%. So now we are down to less than half the number Drum quoted see update above (I sent an email to the study author for clarification but have not heard back. Update: he was nice enough to send me a quick email).
So let's accept this
0.45% 0.7% number for a moment. If GDP started somewhere around 16 trillion in 2010, if we apply a 0.45% the quarterly growth numbers from his chart, we get an incremental economic activity from 2011 through 2013:Q2 of about $333 billion.
So now look at the spending side. The source says that discretionary spending fell by about 2% of GDP over this period. From the graph above, it seems to bite pretty early, but we will assume it fell 1/12 of this 2% figure each quarter, so that by the end of 2013 or beginning of 2014 we get a fall in spending by 2% of GDP. Cumulatively, this would be a reduction in spending over the 2.5 years vs. some "non-austere" benchmark of $388 billion.
Thus, in exchange for running up
$677 billion $388 billion in additional debt, we would have had $445 billion $333 billion in incremental economic activity. A couple of reactions:
- Having the government borrow money and spend it definitely increases near-term GDP. No one disputes that. It is not even in question. Those of us who favor reigning in government spending acknowledge this. The question is, at what cost in terms of future obligations. In fact, this very study Drum is quoting says
Economists agree that failure to shrink prospective deficits and debt will bestow significant economic consequences and risks on future generations. Federal deficits drive up interest rates, “crowding out” private investment. If government borrowing supports consumption (e.g., through Social Security and major health programs) rather than public investment, the nation’s overall capital stock declines, undermining our standard of living. The process is slow but the eventual impact is large.2 In addition, accumulating debt raises the risk of a fiscal crisis. No one can say when this might occur but, unlike crowding out, a debt crisis could develop unexpectedly once debt reached high levels.
High deficits and debt also undermine the efficacy of macroeconomic policies and reduce policymakers’ flexibility to respond to unexpected events. For example, in a recession, it would be harder to provide fiscal stimulus if deficits and debt already were high. Furthermore, fiscal stimulus might be less effective then. Additional deficit spending could be seen as pushing the nation closer to crisis, thereby forcing up interest rates and undercutting the effects of the stimulus. With fiscal policy hamstrung, the burden of counter-cyclical policy is thrust on the Federal Open Market Committee (FOMC) but, particularly in a low interest-rate environment, the FOMC may be unable (or unwilling) to provide additional monetary
- I guess we have pretty much given up on the >1 multiplier, huh? Beggaring our children for incremental economic growth today is a risky enough strategy, but particularly so with the implied
.66.85 multiplier here.