TEDGlobal 2012
Keith Chen: Could your language affect your ability to save money?
Filmed:
Readability: 4.3
1,880,497 views
What can economists learn from linguists? Behavioral economist Keith Chen introduces a fascinating pattern from his research: that languages without a concept for the future -- "It rain tomorrow," instead of "It will rain tomorrow" -- correlate strongly with high savings rates.
Keith Chen - Behavioral economist
Keith Chen's research suggests that the language you speak may impact the way you think about your future. Full bio
Keith Chen's research suggests that the language you speak may impact the way you think about your future. Full bio
Double-click the English transcript below to play the video.
00:16
The global economic financial crisis has reignited public interest
0
500
4500
00:20
in something that's actually one of the oldest questions in economics,
1
5000
3416
00:24
dating back to at least before Adam Smith.
2
8416
2684
00:27
And that is, why is it that countries with seemingly similar economies and institutions
3
11100
5484
00:32
can display radically different savings behavior?
4
16584
3249
00:35
Now, many brilliant economists have spent their entire lives working on this question,
5
19833
4534
00:40
and as a field we've made a tremendous amount of headway
6
24367
3417
00:43
and we understand a lot about this.
7
27784
2433
00:46
What I'm here to talk with you about today is an intriguing new hypothesis
8
30217
3634
00:49
and some surprisingly powerful new findings that I've been working on
9
33851
4032
00:53
about the link between the structure of the language you speak
10
37883
4706
00:58
and how you find yourself with the propensity to save.
11
42604
4396
01:02
Let me tell you a little bit about savings rates, a little bit about language,
12
47000
3067
01:05
and then I'll draw that connection.
13
50067
2350
01:08
Let's start by thinking about the member countries of the OECD,
14
52417
4567
01:12
or the Organization of Economic Cooperation and Development.
15
56984
3301
01:16
OECD countries, by and large, you should think about these
16
60285
3899
01:20
as the richest, most industrialized countries in the world.
17
64184
2638
01:22
And by joining the OECD, they were affirming a common commitment
18
66822
4050
01:26
to democracy, open markets and free trade.
19
70872
3438
01:30
Despite all of these similarities, we see huge differences in savings behavior.
20
74310
4685
01:34
So all the way over on the left of this graph,
21
78995
2450
01:37
what you see is many OECD countries saving over a quarter of their GDP every year,
22
81445
4734
01:42
and some OECD countries saving over a third of their GDP per year.
23
86179
4681
01:46
Holding down the right flank of the OECD, all the way on the other side, is Greece.
24
90860
4768
01:51
And what you can see is that over the last 25 years,
25
95628
3416
01:54
Greece has barely managed to save more than 10 percent of their GDP.
26
99044
3900
01:58
It should be noted, of course, that the United States and the U.K. are the next in line.
27
102944
6874
02:05
Now that we see these huge differences in savings rates,
28
109818
2578
02:08
how is it possible that language might have something to do with these differences?
29
112396
3666
02:11
Let me tell you a little bit about how languages fundamentally differ.
30
116062
3049
02:15
Linguists and cognitive scientists have been exploring this question for many years now.
31
119111
5567
02:20
And then I'll draw the connection between these two behaviors.
32
124678
4610
02:25
Many of you have probably already noticed that I'm Chinese.
33
129288
2608
02:27
I grew up in the Midwest of the United States.
34
131896
2965
02:30
And something I realized quite early on
35
134861
2451
02:33
was that the Chinese language forced me to speak about and --
36
137312
3591
02:36
in fact, more fundamentally than that --
37
140903
2891
02:39
ever so slightly forced me to think about family in very different ways.
38
143794
4090
02:43
Now, how might that be? Let me give you an example.
39
147884
2077
02:45
Suppose I were talking with you and I was introducing you to my uncle.
40
149961
4402
02:50
You understood exactly what I just said in English.
41
154363
2866
02:53
If we were speaking Mandarin Chinese with each other, though,
42
157229
2950
02:56
I wouldn't have that luxury.
43
160179
2066
02:58
I wouldn't have been able to convey so little information.
44
162245
2833
03:00
What my language would have forced me to do,
45
165078
2384
03:03
instead of just telling you, "This is my uncle,"
46
167462
2000
03:05
is to tell you a tremendous amount of additional information.
47
169462
3282
03:08
My language would force me to tell you
48
172744
1819
03:10
whether or not this was an uncle on my mother's side or my father's side,
49
174563
3416
03:13
whether this was an uncle by marriage or by birth,
50
177979
3084
03:16
and if this man was my father's brother,
51
181063
2232
03:19
whether he was older than or younger than my father.
52
183295
2784
03:21
All of this information is obligatory. Chinese doesn't let me ignore it.
53
186079
4133
03:26
And in fact, if I want to speak correctly,
54
190212
2166
03:28
Chinese forces me to constantly think about it.
55
192378
3117
03:31
Now, that fascinated me endlessly as a child,
56
195495
3949
03:35
but what fascinates me even more today as an economist
57
199444
3235
03:38
is that some of these same differences carry through to how languages speak about time.
58
202679
5301
03:43
So for example, if I'm speaking in English, I have to speak grammatically differently
59
207980
4249
03:48
if I'm talking about past rain, "It rained yesterday,"
60
212229
2733
03:50
current rain, "It is raining now,"
61
214962
2232
03:53
or future rain, "It will rain tomorrow."
62
217194
2434
03:55
Notice that English requires a lot more information with respect to the timing of events.
63
219628
4827
04:00
Why? Because I have to consider that
64
224455
2007
04:02
and I have to modify what I'm saying to say, "It will rain," or "It's going to rain."
65
226462
4783
04:07
It's simply not permissible in English to say, "It rain tomorrow."
66
231245
4116
04:11
In contrast to that, that's almost exactly what you would say in Chinese.
67
235361
4184
04:15
A Chinese speaker can basically say something
68
239545
2316
04:17
that sounds very strange to an English speaker's ears.
69
241861
2584
04:20
They can say, "Yesterday it rain," "Now it rain," "Tomorrow it rain."
70
244445
4567
04:24
In some deep sense, Chinese doesn't divide up the time spectrum
71
249012
3863
04:28
in the same way that English forces us to constantly do in order to speak correctly.
72
252875
6433
04:35
Is this difference in languages
73
259308
1584
04:36
only between very, very distantly related languages, like English and Chinese?
74
260892
4195
04:40
Actually, no.
75
265087
958
04:41
So many of you know, in this room, that English is a Germanic language.
76
266045
3733
04:45
What you may not have realized is that English is actually an outlier.
77
269778
3915
04:49
It is the only Germanic language that requires this.
78
273693
3250
04:52
For example, most other Germanic language speakers
79
276943
2884
04:55
feel completely comfortable talking about rain tomorrow
80
279827
3017
04:58
by saying, "Morgen regnet es,"
81
282844
1966
05:00
quite literally to an English ear, "It rain tomorrow."
82
284810
3900
05:04
This led me, as a behavioral economist, to an intriguing hypothesis.
83
288710
5067
05:09
Could how you speak about time, could how your language forces you to think about time,
84
293777
4249
05:13
affect your propensity to behave across time?
85
298026
3751
05:17
You speak English, a futured language.
86
301777
2900
05:20
And what that means is that every time you discuss the future,
87
304677
3167
05:23
or any kind of a future event,
88
307844
1567
05:25
grammatically you're forced to cleave that from the present
89
309411
3399
05:28
and treat it as if it's something viscerally different.
90
312810
2633
05:31
Now suppose that that visceral difference
91
315443
2500
05:33
makes you subtly dissociate the future from the present every time you speak.
92
317943
4200
05:38
If that's true and it makes the future feel
93
322143
1913
05:39
like something more distant and more different from the present,
94
324056
2983
05:42
that's going to make it harder to save.
95
327039
2654
05:45
If, on the other hand, you speak a futureless language,
96
329693
2551
05:48
the present and the future, you speak about them identically.
97
332244
3383
05:51
If that subtly nudges you to feel about them identically,
98
335627
2983
05:54
that's going to make it easier to save.
99
338610
2384
05:56
Now this is a fanciful theory.
100
340994
2550
05:59
I'm a professor, I get paid to have fanciful theories.
101
343544
2900
06:02
But how would you actually go about testing such a theory?
102
346444
4234
06:06
Well, what I did with that was to access the linguistics literature.
103
350678
4066
06:10
And interestingly enough, there are pockets of futureless language speakers
104
354744
4383
06:15
situated all over the world.
105
359127
1933
06:16
This is a pocket of futureless language speakers in Northern Europe.
106
361060
3366
06:20
Interestingly enough, when you start to crank the data,
107
364426
2901
06:23
these pockets of futureless language speakers all around the world
108
367327
3233
06:26
turn out to be, by and large, some of the world's best savers.
109
370560
3934
06:30
Just to give you a hint of that,
110
374494
2166
06:32
let's look back at that OECD graph that we were talking about.
111
376660
2750
06:35
What you see is that these bars are systematically taller
112
379410
3384
06:38
and systematically shifted to the left
113
382794
2132
06:40
compared to these bars which are the members of the OECD that speak futured languages.
114
384926
4518
06:45
What is the average difference here?
115
389444
1463
06:46
Five percentage points of your GDP saved per year.
116
390907
3286
06:50
Over 25 years that has huge long-run effects on the wealth of your nation.
117
394193
4734
06:54
Now while these findings are suggestive,
118
398927
2700
06:57
countries can be different in so many different ways
119
401627
2066
06:59
that it's very, very difficult sometimes to account for all of these possible differences.
120
403693
4384
07:03
What I'm going to show you, though, is something that I've been engaging in for a year,
121
408077
4032
07:08
which is trying to gather all of the largest datasets
122
412109
2323
07:10
that we have access to as economists,
123
414432
2292
07:12
and I'm going to try and strip away all of those possible differences,
124
416724
3382
07:16
hoping to get this relationship to break.
125
420106
2654
07:18
And just in summary, no matter how far I push this, I can't get it to break.
126
422760
5031
07:23
Let me show you how far you can do that.
127
427791
1765
07:25
One way to imagine that is I gather large datasets from around the world.
128
429556
4633
07:30
So for example, there is the Survey of Health, [Aging] and Retirement in Europe.
129
434189
3734
07:33
From this dataset you actually learn that retired European families
130
437923
3834
07:37
are extremely patient with survey takers.
131
441757
2633
07:40
(Laughter)
132
444390
1916
07:42
So imagine that you're a retired household in Belgium and someone comes to your front door.
133
446306
4384
07:46
"Excuse me, would you mind if I peruse your stock portfolio?
134
450690
4584
07:51
Do you happen to know how much your house is worth? Do you mind telling me?
135
455274
3532
07:54
Would you happen to have a hallway that's more than 10 meters long?
136
458806
3267
07:57
If you do, would you mind if I timed how long it took you to walk down that hallway?
137
462073
4501
08:02
Would you mind squeezing as hard as you can, in your dominant hand, this device
138
466574
3897
08:06
so I can measure your grip strength?
139
470471
1512
08:07
How about blowing into this tube so I can measure your lung capacity?"
140
471983
4023
08:11
The survey takes over a day.
141
476006
2884
08:14
(Laughter)
142
478890
1483
08:16
Combine that with a Demographic and Health Survey
143
480373
3900
08:20
collected by USAID in developing countries in Africa, for example,
144
484273
4450
08:24
which that survey actually can go so far as to directly measure the HIV status
145
488723
5151
08:29
of families living in, for example, rural Nigeria.
146
493874
3099
08:32
Combine that with a world value survey,
147
496973
1901
08:34
which measures the political opinions and, fortunately for me, the savings behaviors
148
498874
4433
08:39
of millions of families in hundreds of countries around the world.
149
503307
4699
08:43
Take all of that data, combine it, and this map is what you get.
150
508006
3818
08:47
What you find is nine countries around the world
151
511824
2250
08:49
that have significant native populations
152
514074
2652
08:52
which speak both futureless and futured languages.
153
516726
4047
08:56
And what I'm going to do is form statistical matched pairs
154
520773
3534
09:00
between families that are nearly identical on every dimension that I can measure,
155
524307
5593
09:05
and then I'm going to explore whether or not the link between language and savings holds
156
529900
3537
09:09
even after controlling for all of these levels.
157
533437
3483
09:12
What are the characteristics we can control for?
158
536920
2217
09:15
Well I'm going to match families on country of birth and residence,
159
539137
2764
09:17
the demographics -- what sex, their age --
160
541901
2402
09:20
their income level within their own country,
161
544303
2134
09:22
their educational achievement, a lot about their family structure.
162
546437
3015
09:25
It turns out there are six different ways to be married in Europe.
163
549452
3518
09:28
And most granularly, I break them down by religion
164
552970
4200
09:33
where there are 72 categories of religions in the world --
165
557170
3315
09:36
so an extreme level of granularity.
166
560485
1717
09:38
There are 1.4 billion different ways that a family can find itself.
167
562202
4533
09:42
Now effectively everything I'm going to tell you from now on
168
566735
4049
09:46
is only comparing these basically nearly identical families.
169
570784
3050
09:49
It's getting as close as possible to the thought experiment
170
573834
2500
09:52
of finding two families both of whom live in Brussels
171
576334
2933
09:55
who are identical on every single one of these dimensions,
172
579267
3000
09:58
but one of whom speaks Flemish and one of whom speaks French;
173
582267
3116
10:01
or two families that live in a rural district in Nigeria,
174
585383
2717
10:04
one of whom speaks Hausa and one of whom speaks Igbo.
175
588100
3833
10:07
Now even after all of this granular level of control,
176
591933
3884
10:11
do futureless language speakers seem to save more?
177
595817
3166
10:14
Yes, futureless language speakers, even after this level of control,
178
598983
3653
10:18
are 30 percent more likely to report having saved in any given year.
179
602636
3698
10:22
Does this have cumulative effects?
180
606334
1816
10:24
Yes, by the time they retire, futureless language speakers, holding constant their income,
181
608150
4420
10:28
are going to retire with 25 percent more in savings.
182
612570
3068
10:31
Can we push this data even further?
183
615638
2481
10:34
Yes, because I just told you, we actually collect a lot of health data as economists.
184
618119
5299
10:39
Now how can we think about health behaviors to think about savings?
185
623418
3883
10:43
Well, think about smoking, for example.
186
627301
2833
10:46
Smoking is in some deep sense negative savings.
187
630134
3183
10:49
If savings is current pain in exchange for future pleasure,
188
633317
3666
10:52
smoking is just the opposite.
189
636983
1308
10:54
It's current pleasure in exchange for future pain.
190
638291
2859
10:57
What we should expect then is the opposite effect.
191
641150
2950
11:00
And that's exactly what we find.
192
644100
1768
11:01
Futureless language speakers are 20 to 24 percent less likely
193
645868
3767
11:05
to be smoking at any given point in time compared to identical families,
194
649635
3415
11:08
and they're going to be 13 to 17 percent less likely
195
653050
2901
11:11
to be obese by the time they retire,
196
655951
2217
11:14
and they're going to report being 21 percent more likely
197
658168
2463
11:16
to have used a condom in their last sexual encounter.
198
660631
2287
11:18
I could go on and on with the list of differences that you can find.
199
662918
3483
11:22
It's almost impossible not to find a savings behavior
200
666401
3800
11:26
for which this strong effect isn't present.
201
670201
2599
11:28
My linguistics and economics colleagues at Yale and I are just starting to do this work
202
672800
4750
11:33
and really explore and understand the ways that these subtle nudges
203
677550
5167
11:38
cause us to think more or less about the future every single time we speak.
204
682717
5395
11:44
Ultimately, the goal,
205
688112
2301
11:46
once we understand how these subtle effects can change our decision making,
206
690413
4199
11:50
we want to be able to provide people tools
207
694612
2950
11:53
so that they can consciously make themselves better savers
208
697562
2808
11:56
and more conscious investors in their own future.
209
700370
3259
11:59
Thank you very much.
210
703629
2267
12:01
(Applause)
211
705896
6368
ABOUT THE SPEAKER
Keith Chen - Behavioral economistKeith Chen's research suggests that the language you speak may impact the way you think about your future.
Why you should listen
Does the future look like a different world to you, or more like an extension of the present? In an intriguing piece of research, Keith Chen suggests that your attitude about the future has a strong relationship to the language you speak. In a nutshell, some languages refer to the future using verb helpers like "will" and "shall," while others don't have specific verbs to refer to future actions. Chen correlated these two different language types with remarkably different rates of saving for the future (guess who saves more?). He calls this connection the "futurity" of languages. The paper is in the process of being published by the American Economic Review, and it's already generated discussion. Chen says: "While the data I analyze don’t allow me to completely understand what role language plays in these relationships, they suggest that there is something really remarkable to be explained about the interaction of language and economic decision-making. These correlations are so strong and survive such an aggressive set of controls, that the chances they arise by random lies somewhere between one in 10,000 and one in 10^32."
Chen excels in asking unusual questions to yield original results. Another work (with Yale colleague and TEDGlobal 2009 speaker Laurie Santos) examined how monkeys view economic risk--with surprisingly humanlike irrationality. While a working paper asks a surprising, if rhetorical, question: Does it make economic sense for a woman to become a physician?
Chen is currently Uber's Head of Economic Research and is an associate professor of economics at UCLA .
Keith Chen | Speaker | TED.com