Bryan Caplan  

Complacent Trust

UBI and Health Care: What's Wr... Henderson on NAFTA and Mercant...
Together with Zach Weinersmith, I'm currently writing All Roads Lead to Open Borders, my non-fiction graphic novel on the philosophy and social science of immigration.  While there are many differences between a non-fiction graphic novel and a traditional work of non-fiction, I still strive to read broadly and deeply before I write.  Now, for the chapter on "Crimes Against Culture," I'm immersed in the vast academic literature on trust

In economics, the word on the street is that trust is good for growth.  And from what I've seen, the leading papers confirm this result.  What's striking, though, is that lower-profile papers occasionally reach rather different conclusions.  This wouldn't be noteworthy if the lower-profile papers were clearly lower-quality papers.  But since the econometrics are simple, the main difference seems to be that the lower-profile papers were published later using more extensive data.  In academia, the early bird gets the worm, even if the later bird is closer to the truth.

In coming months, I'll be blogging some contrarian papers on trust.  Today, let's start with Felix Roth's "Does Too Much Trust Hamper Economic Growth?" (Kyklos, 2009).  Big result: Estimating growth as a function of trust and trust squared yields a strong quadratic relationship:
Regression 4, taking a country sample without transition countries, modulates trust as a curvilinear relationship to economic growth by including the squared term of interpersonal trust into the regression. Astonishingly, the curvilinear relationship is highly significant. All variables in the regression have the expected signs and are highly significant (99% level of significance). The linear and squared term of interpersonal trust are each statistically significant: 0.16 (4.42) and -0.0015 (-3.24). These estimates imply that starting from a low-trust country (where the interpersonal trust value is for instance 2.8, as in Brazil), increases in interpersonal trust tend to stimulate economic growth. However, the positive influence attenuates as the level of trust rises and reaches zero when the indicator takes on a mid-range of 53.3. Therefore, an increase in the level of trust appears to enhance economic growth in countries that have initial low levels of trust but to retard economic growth for countries that have already achieved a substantial level of trust.
Within the OECD sub-sample, "The positive influence attenuates as the level of trust rises and reaches zero when the indicator takes on a mid-range of 42.5."  Here's the cool graph:
trust2.jpgWith fixed effects, moreover, the linear effect of trust on growth looks negative:
[T]aking panel data and using a fixed-effects estimation for a 41-country sample over the time period from 1980 to 2004 and with a total of 129 observations, the paper points out that economic growth is negatively related to an increase in trust. This negative finding is in contrast to most empirical findings using a cross-sectional design. The negative relationship seems to be mainly driven by developed countries from the OECD (here specifically Poland, Greece, and the United States), and the EU-15 (here particularly the
United Kingdomand Finland), and very strongly by LMEs [liberal market economies] and Scandinavian countries.
But how on Earth could trust ever be bad?  The most obvious story is Cowenian.  Excessive trust leads to (or is perhaps identical with) complacency.  If everything is awesome, who needs innovation and dynamism?

Is this paper the final word?  Of course not.  Maybe the pro-trust consensus is right and Roth is wrong.  But given the prevalence of academic groupthink, the opposite wouldn't surprise me.  At all.

COMMENTS (12 to date)
Andrew Hofer writes:

My Cousin, Charles Heckscher, has written on the subject. He is not contrarian, but comes from very different priors than yourself. Very statist but also highly consequentialist.

Intuitionist writes:

What happens when we include demographic growth as a confounding variable?

Plucky writes:

(Note: I'm too lazy to read the paper, so feel free to discard if I'm way off-base)
(Note 2: The link to the paper goes to a previous post of yours, not the paper)
Several comments:

1) At the pure visual level, a lot of the the negative-2nd order seems to be driven by a) Korea (ko?) as a mid-trust, high-growth semi-outlier and b) pt (Portugal?) in the low-trust/low-growth semi-outlier. If the fit is appreciably different without those then it's probably somewhat suspect. Arbitrarily throwing them out isn't exactly good practice either, but if 2 data points are (mathematically) responsible for much of the result, then one should not be strongly confident in the result. Also, it woudl be interesting to see what a piece-wise linear model would look like for this data instead of a quadratic model

2) A useful exercise for considering the possibility of mixed/conflicting/changing marginal effects of trust on growth would doing a Solow-style split of the growth into capital, labor, and TFP components, and then estimating the effects of trust on the component level. One could hypothesize:
a) Trust should be positive at all margins for capital-driven growth, since trust would generally indicate legal, social, political, and rule-of-law stability
b) Trust could have positive, negative, or no effect on the labor contribution for a variety of reasons (more on this below)
c) Trust should have a positive effect on TFP since it probably indicates a low level of drag and inefficiency arising from disputes. This effect could very plausibly be subject to diminishing returns at high trust levels

elaboration of b: Since in Solow-style growth accounting, the labor contribution is simply total amount of labor (as opposed to quality of labor, which would show up in TFP), there are plenty of possible effects, some of which would result in a negative correlation of trust to growth. If a country maintained high trust by having a restrictive immigration policy to maintain cultural or ethnic homogeneity, then the size of the labor force would grow more slowly and thus be negative for overall growth. In those cases, the labor force of a high-trust country would just be a function of its birth rate and health levels (all the countries in the high-trust / low growth cluster have had sub-replacement fertility for many years. Attempting to establish any kind of causal relationship is probably impossible with macro data) . On the other hand, if high-trust countries were especially attractive for immigrants, good at integrating them, and accommodated that, then you might get a positive L result.

Also, in a Solow-style growth-accounting exercise, a lot of the growth in L in developed countries during the period in question has come from increased female participation in the labor force. Trying to establish any link between that and societal trust levels is also probably impossible to do with macro data, so ideally a trust-growth link should hold whether or not increased female labor force participation is considered exogenous or endogenous

jj writes:

I see dk a lot at the top end of trust. Does higher trust allow a larger welfare state, lowering growth?

Tiago writes:

Shouldn't they have examined the relationship of trust and GDP per capita instead of growth? If countries are all near their steady state gdp growth could come from a global increase in technology. In this case, the relationship between growth and trust could be zero even if trust greatly affects GDP per capita levels by increasing productivity.

Mr. Econotarian writes:

Too much trust implies too much trust in government.

You have to trust government enough to know that it will follow the rule of law, and won't come steal your property for no reason.

But if you trust it too much, you won't care when it picks your pocket, because you're sure that money is going to something useful, right?

Many of those high-trust countries may also be racially/culturally homogenous, which defuses anti-socialist racism of the sort "no welfare for [lazy/criminal] [outgoup]".

Sam Hardwick writes:

One major driver for striving in the workplace is to be able to afford a nice area to live. When trust is high, more places are nice, so you don't need to outdo yourself that much.

Pajser writes:

Poverty => crime, corruption, nepotism => lower trust
Poverty => faster growth in convergence period, starting in 1980's

Michael writes:

I'm sorry, but this is rather a cavalier assumption on the direction of causation. It probably works both ways, but what the paper measures is clearly the known fact that development
a) increases trust
b) eventually reduces growth

Nothing unintuitive about this at all: relative growth rates are much higher when levels are low and catch up happens. As it happens, trust increases as levels increase

Bottom line: growth rates are a good target for policy, but are often the wrong measure of longer-term economic success, where you want to capture the cumulated effects

Noah Carl writes:

Most of the over-time variance in generalized trust at the country level is measurement error, so a panel-analysis is fairly pointless.

Ben writes:

Doesn't include growth in GDP per capita. I'd imagine Nigeria's or India's populations are growing in percentage terms a hell of a lot faster than Sweden's or Norway's.

What's more likely is that:
More trust->Better Institutions->higher incomes->less 'room' for growth (as growth is obviously harder for a high-income country more reliant on ideas than more physical capital)-> slower growth rates

David Belkin writes:

If you haven't already, you'll want to dig into
Karen S. Cook, Russell Hardin, and Margaret Levi, Cooperation Without Trust? (Russell Sage, 2007).


Return to top