Monday, October 12, 2009

How Well Known Was the Productivty Surge of the Early 2000s?

In my previous post I noted that the rebound in productivity growth that started in the 1990s did not end with the tech bubble bursting in 2000. Rather, it took a temporary respite and then continued to accelerate for a few more years. The point of my sharing this information is that the robust productivity gains of the early-to-mid 2000s imply interest rates should have been higher. Instead they were dropping, a sign that monetary policy was too loose. This development also sheds light on the origins of the deflation scare of 2003: inflation was falling because of positive aggregate supply shocks (i.e. the rapid productivity gains) not negative aggregate demand shocks as was believed back then. Nominal spending, in fact, was soaring during this period. Consequently, the U.S. economy was getting juiced-up on easy money at same time it was being buffeted with positive productivity shocks. This policy response primed the U.S. economy for the emergence of economic imbalances.

A few commentators objected to this interpretation of events. They conceded that the actual productivity growth rate may have continued to surge, but questioned whether productivity expectations kept up with reality. Instead, they surmised that expectations of productivity growth declined after the tech bubble burst. If so, the expected marginal product of capital would have declined, investment demand would have decreased, and the neutral rate of interest also would have fallen. This is a great objection and one I had to check out. First, I went to the Survey of Economic Forecasters from the Philadelphia Fed and looked at the forecast of the average productivity growth rate over the next 10 years. This series begins in 1992 and is graphed below (click on figure to enlarge):


This figure show that the 10-year productivity forecast decline slightly in 2002 but resumed its climb in 2003. It peaks out in 2004, with mild declines in 2005 and 2006. By 2007 the writing was on the wall and the forecast begins to rapidly decline. This figure suggest, then, the actual productivity gains during this time were known and shaping the long-term forecast of productivity growth.

As a robustness check--and because I vaguely remembered there being a lot productivity stories in the media back in 2003-2004--I went to Lexus-Nexus and did a U.S. newspaper and news wire search with the following key words: productivity growth, accelerated or increased or pick up or faster or miracle or upward or improved or strong or robust or sustained. This search was intended to pick up articles, if any, discussing the productivity surge at the time. Here is what I found over 1992-2009 (click on figure to enlarge):

This figure shows a similar pattern: news stories on positive productivity growth articles temporarily declined in 2001 but then continued their upward momentum thereafter. The series peaks in 2005 and then begins a dramatic collapse. The positive productivity news stories bubble pops at that time.

The above figures may not convince everyone, but they are enough evidence for me to conclude that my earlier interpretation of events during the early-to-mid 2000s cannot be too far off the mark.

22 comments:

  1. Since GDP growth itself for the period was illusory , because it was "borrowed" from the current and future years in the form of crushing consumer debt , any boasting about productivity for that period is ludicrous.

    Send 20 grand to every middle-class family tomorrow , and I guarantee you'll generate a spike in GDP , and , of course , productivity , but it won't say a thing about our productive capacity.

    Nice try , though.

    ReplyDelete
  2. Anonymous:

    Total factor productivity (TFP)-- not just labor productivity--was surging during this period. Even utilization adjusted TFP shows rapid gains (See the data here)Robert Gordon has shown the productivity gains were real. Yes, there was a borrowing binge but that does not mean there was no productivity surge.

    ReplyDelete
  3. Productivity gains were real, but I don't know that they were indicative of what you say they are.

    Productivity growth during tepid, jobless recoveries hardly reflects high returns on marginal capital investments. I would argue that the current surge in productivity (hours worked falling and production rising) also reflects this little-understood dynamic. Does strong 3q09 productivity reflect high marginal investment returns? Unlikely given all the excess capacity in the economy. This is the nub of the current problem in the economy: if productivity growth does not spur investment, then you get downward pressure on Personal Income with little gain from Business Investment. This is ultimately deflationary.

    ReplyDelete
  4. David,

    Robert Gordon makes the case in this paper that the productivity gains were real and based on (1) aggressive corporate cost cutting and (2) the lagged effect of the massive investments in information and communication technology in the late 1990s. I would add to his list the continued opening up of Asia and as another productivity booster for the United States (though this may be implicit in his (1) above).

    ReplyDelete
  5. Have you looked at Laura Veldkamp's work? Strikes me it might be of interest to you, especially her paper in JME about learning asymmetries in real business cycles. Learning about productivity growth plays an important role.

    ReplyDelete
  6. Oh yes, I also remember that Paul Beaudry had a paper in AER something like "News, stock prices and economic fluctuations" It again makes the case for TFP news as crucial in business cycle dynamics.

    ReplyDelete
  7. Sorry to hog the comments, but the RBC people, some of them, do see the current episode as another data point in their favor. This time the TFP innovation was in the financial sector, boosting the economy as we learned about the productivity growth in securitization. In this interpretation, low interest rates and monetary policy are irrelevant, its all due to learning about TFP.

    ReplyDelete
  8. I think this is a nice follow on in thinking about how to observe expected productivity as opposed to actual productivity. The main reason it doesn't persuade me much is the nature of productivity forecasting compared to the kinds of expectations that actually determine investment demand. An economic forecaster probably frames inflation expectations somewhat similarly to a typical household, but I'm not sure businesses asking themselves whether good projects are available feels a lot like a productivity forecast.

    More tangibly, the Philly forecast shows only a small blip in its generally high productivity forecasts throughout the early 2000s. Yet private fixed nonresidential investment was very weak from the end of 2000 through early 2003, the period where real interest rates declined. Even in 2004 - 2006, it grew at lower rates than it had at any year between 1993-2000. And, as you would expect with too-low productivity forecasts, profits went on to consistently surprise on the upside.

    ReplyDelete
  9. "Robert Gordon makes the case in this paper that the productivity gains were real and based on (1) aggressive corporate cost cutting and (2) the lagged effect of the massive investments in information and communication technology in the late 1990s. I would add to his list the continued opening up of Asia and as another productivity booster for the United States (though this may be implicit in his (1) above)."

    Yes there was certainly cost-cutting. Pricing power was nil even with the surge in debt accumulation by consumers ; the end was near , and everyone , consumers and corporate heads alike , sensed it. Thus , no capital investment but instead cut costs to the bone , and , voila! , productivity gains.

    Yes , there were "lagged" effects in ICT contributions , but they were lagged in the same way a plowed field that lays fallow will show "lagged" productivity gains when the farmer can afford to plant the field after he gets a big home equity draw on his new NINJA loan.

    It was all a mirage. We faced the same situation in the early 2000s that we do now --- a demand-constrained economy resulting from the disenfranchisement of working-class Americans. And , again, you can re-create the phony productivity by sending out big checks to middle-income workers. Corporations will continue to cut costs , because they know it won't last. Profits and productivity measures will rise , but they will be meaningless forecasting tools , because you're just postponing the reckoning day.

    Your "benign deflation" of the early 2000s was no different qualitatively than the deflation of today , just like Japan's is no different than it was five years ago.

    You can rewrite economic history by juggling fudge factors , certainly , but you can't make the real economy respond to its new "life story". It will respond to the facts as they truly exist , and if you want to see this economy become strong again , you need to acknowledge the forces that made it so in the past , and also acknowledge those forces that made it much less strong in the more recent past.

    That sort of acknowledgement is difficult to extract , however , from those who are paid to rewrite history in an effort to resuscitate a failed ideology , so I won't look for here.

    ReplyDelete
  10. Please read grandfathers economic report to understand how productivity numbers could have been fudged.

    http://mwhodges.home.att.net/

    God bless.

    ReplyDelete
  11. I happened to read anonymous's latest comments just after reading Paul Romer on Elinor Ostrom's Nobel and it seems anonymous may have a point. Ostrom's research is about how and why people follow or dont follow the rules of the game. The usual econ story about Pareto optimality of competitive equilibria assumes that people follow the rules. The events of the last 10 years have dealt severe blows to trust in the system. US social capital has been badly eroded as we see episode after episode of corporate fraud and accounting fraud. This does not show up immediately in TFP but you can bet that in the long run it could be disastrous for US outlook. Going down the path of Argentina does not look impossible.

    ReplyDelete
  12. ECB said :

    "Going down the path of Argentina does not look impossible."

    Too late , we're already there. See the latest by Saez et al and compare Argentina and the U.S.

    http://elsa.berkeley.edu/~saez/atkinson-piketty-saezNBER09topincomes.pdf

    ReplyDelete
  13. Anonymous,

    Where is your evidence? I have am trying to be open to your view, but quite frankly with your emotionally-charged claims its hard to know what to take seriously.

    And your lagged effect analogies fail to convince me. For example, you state the following:

    "Yes , there were "lagged" effects in ICT contributions , but they were lagged in the same way a plowed field that lays fallow will show "lagged" productivity gains when the farmer can afford to plant the field after he gets a big home equity draw on his new NINJA loan."

    There is a world of difference between a fallow field that meets a large loan and the large investments in information and communication technology in the late 1990s whose benefits took some time to be realized. Speak directly to delayed IT returns story--why doesn't it make sense?

    ReplyDelete
  14. David,

    I don't dispute that productivity grew. I dispute that rising productivity signaled attractive or rising marginal returns on investment. Again, I believe we face the same dynamic today, except without housing as a viable target for marginal investment capital. That dynamic -- pressure on wages combined with anemic investment -- is inherently deflationary.

    ReplyDelete
  15. "Yes , there were "lagged" effects in ICT contributions , but they were lagged in the same way a plowed field that lays fallow will show "lagged" productivity gains when the farmer can afford to plant the field after he gets a big home equity draw on his new NINJA loan."

    you answered:

    "There is a world of difference between a fallow field that meets a large loan and the large investments in information and communication technology in the late 1990s whose benefits took some time to be realized. Speak directly to delayed IT returns story--why doesn't it make sense?"

    Just because there is a world of difference doesn't make it a bad analogy. When I say: " Discerning Beckworth's agenda is 'As Easy as Falling Off a Log' " , its still a good analogy , even though no logs are involved in your agenda.

    I'll try to be more direct , however. What if the ICT enhancements were essentially fully-formed by 2001 , say , but a demand-constrained consumer market kept those productivity-enhancing features from being realized ( think : fallow field ). Then , the miracles of no-doc loans supercharged consumer purchases , allowing the enhanced productivity to reveal itself ( think : new seeds growing into new crops ).

    The spike , and subsequent decline , in productivity you point to curiously coincides with the spike and fall in liquidity to consumers. That's all I'm saying. I wouldn't go so far as to contort reams of data to prove that I'm correct , as Gordon does , but I do think it stands up well as an explanation that should have been disproved by your side before claiming that the world has unfairly overlooked Bush's "Productivity Surge ".

    ReplyDelete
  16. Anonymous:

    Thanks for the clarification. I don't deny that ample liquidity conditions coincided with the robust productivity numbers. I just put a slightly different interpretation on what happened. The Fed allowed liquidity conditions to surge because they saw no pick up in inflation. The stable inflation, in turn, was the result of the productivity surge. Stated differently, the Fed saw the productivity gains as a chance to get inflation-free monetary easing. Stable inflation, however, did not mean macro stability in this case since the productivity gains behind the low inflation implied interest rates should have been higher and liquidity conditions tighter than they were.

    My agenda here is simple: make the Fed more effective in stabilizing nominal spending. I believed it failed on the upside in 2003-2005 (when it took advantage of the productivity gains to get inflation-free monetary easing and, as result, stimulated a nominal spending spree) and failed on the downside in late 2008,early 2009 (when it allowed nominal spending to collapse). Thus, I am all for the Fed adopting a nominal income targeting rule. Contrary to your claims, there is nothing more to my agenda. (Just ask ECB who gets sick of my nominal income targeting rants)

    ReplyDelete
  17. Making the Fed more effective is certainly a laudable agenda , and should be easier to do than making it less effective. That would take some serious effort.

    If not your agenda , your ideology has been revealed in previous posts , for example :

    http://macromarketmusings.blogspot.com/2008/04/real-wages-have-kept-up-with.html

    In which M. Feldstein says :

    "The relation between wages and productivity is important because it is a key determinant of the standard of living of the employed population as well as of the distribution of income between labor and capital. If wages rise at the same pace as productivity, labor’s share of national income remains essentially unchanged. This paper presents specific evidence that this has happened: the share of national income going to employees is at approximately the same level now as it was in 1970."

    Your ending comment :


    "I want to believe these findings, but they seem too good to be true. What do you think?"


    The operative phrase being : " I WANT to believe ....".

    Why would you WANT to believe such findings ? Because it means that supply-side works just dandy ? No course change necessary. Bring on more tax cuts for the rich , kill the cap gains and estate ( oops , sorry , DEATH ) taxes , deregulate , offshore jobs. Drill-Baby-Drill. Obama is Hitler , but worse , since he was born in Kenya.

    The Friedman piece was such a transparent effort to obscure the massive redistribution of incomes to the top few percent , but you WANTED to believe it , rather than challenge it.

    Ok , so maybe it's only revealing of your ideology , rather than your agenda. Is that a distinction without a difference ?

    My experience suggests the agenda almost inevitably accompanies the ideology.

    ReplyDelete
  18. Anonymous:

    At least your are reading my blog! If anything, the post to which you refer does reflect my desire to see neoclassical economic theory (e.g. marginal productivity of labor = real wage) work in practice. On that count I am guilty as charged. However, there is definitely a libertarian streak to me though I try to reach conclusions based on the data, even when it runs contrary to my priors. I hope you do the same.

    ReplyDelete
  19. I'll give you credit for this , you allow comments , including highly critical ones. That's more than can be said for many others , like Mankiw , for example.

    "I try to reach conclusions based on the data, even when it runs contrary to my priors"

    You must be highly selective , maybe subconciously , about the data you peruse. Your topic guide titled " Real Wages " , has three entries , all smoke and mirrors of the type I referenced previously.

    I won't be bothering you again. I've seen enough here. I actually do believe your agenda is clear , and aligns with folks like Kudlow. Though I find him repulsive , I respect the fact that he's transparent. To be any more transparent , he'd have to tattoo his forehead with : " Greed , for the lack of a better word , is GOOD ! "

    He's appealing to the MSM red-meat crowd , however , so it makes sense. Here , you're addressing readers with at least a handful of functional neurons. So , the strategy is to shape opinion with more subtlety , as the posts under "Real Wages" might be expected to do. You and Kudlow are both trying to boil the same frog , but you're doing it verrrryyy slowly.

    Look at the Saez data in the link I gave above. Make a prediction about how you think income inequality will correlate with economic growth over the next 3 , 5, or 10 years among the advanced economies. I'll come back then and see how you did.

    ReplyDelete
  20. One additional thought on this. You offered two alternatives to lower expected incremental productivity to explain the lower real long term rates in the early to mid 2000s: A lower term premium as a result of the Great Moderation or expectations of more accommodative policy. Both of these seem hard to swallow for me, given no reduction in productivity expectations.


    For one, could the term premium really have decreased from the good years of the late 1990s to the much more tumultuous world of 2001-2004? The great moderation had seemed even greater and more moderate in the earlier period -- hard to see why the term premium would shrink just as things were getting rocky again. I am definitely on board with the idea that there may have been lulling manifestations from extended low volatility or even the "Greenspan Put," but it seems harder to see how a reduction in the term premium from 2000 to 2003 is a likely candidate.

    And I'm not fully sure what you mean by expectations of more accommodative policy. I think of accommodative policy in relative terms as opposed to absolute interest rates, i.e. "rates will be targeted such that nominal spending will rise faster than the previous path given future money demand." So if the market thought the Fed was going to be accommodative in the future, wouldn't you expect to have seen higher inflation expectations (which of course you didn't) and as opposed to lower expected future spot rates? Clearly there was an expectation of future low spot rates (save for the unlikely possibility of a substantial negative term premium), but does this easily correspond to a projection of accommodative policy, especially in tandem with high productivity expectations?

    I would think that a market that expected high productivity and overly accommodative Fed policy wouldn't produce low interest rates.

    ReplyDelete
  21. dlr:

    Again, you raise a good question: why would the term premium suddenly decline in 2001-2004? I am not sure I have a good answer--hence, my "tentative" explanation--but a number of empirical studies show a decline in the term premium at this time. Maybe the same giddy spirits behind the housing boom were also behind the term premium decline. I have, however, argued before that the "conundrum" that arose later in 2005-2006 could have been the bond market simply pricing in a recession. There is even a study that shows as much. I blogged about it here.

    ReplyDelete
  22. On term premium, theres a paper by JB Durham "Implied interest skew, term premium and the conundrum"
    as well as one by John Cochrane
    available at his website

    ReplyDelete