%%{init: {"flowchart": {"useMaxWidth": true, "nodeSpacing": 90, "rankSpacing": 75, "padding": 48, "wrappingWidth": 700, "curve": "basis"}, "themeVariables": {"fontSize": "20px", "fontFamily": "Verdana, Geneva, sans-serif"}, "markdownAutoWrap": false, "width": 1150, "height": 620}}%%
flowchart LR
S["$1,000 goes missing"] --> A["Someone else gains it"]
S --> B["The road gets thinner"]
B --> C["Journeys get slower"]
B --> D["Traders stay away"]
B --> E["Trust drains away"]
classDef src fill:#1e293b,stroke:#1e293b,color:#f9fafb,stroke-width:2px;
classDef mid fill:#b44527,stroke:#b44527,color:#f9fafb,stroke-width:2px;
classDef out fill:#f9fafb,stroke:#4a7c6f,color:#1e293b,stroke-width:2.5px;
class S src;
class B mid;
class A,C,D,E out;
The Politics of Corruption
Lecture 5: The Costs of Corruption
What does corruption actually cost?
A city under bad government, painted in Siena around 1338.
Ambrogio Lorenzetti, Effects of Bad Government in the City, Palazzo Pubblico, Siena. Public domain, via Wikimedia Commons.
Painters have shown us the cost for 700 years. Can we count it?
A village, a road, and a missing quarter
- A village in Java receives about US$8,800 to surface its road
- The village reports, line by line, what it spent
- Engineers later dig into the road and measure what is actually there
- About a quarter of the reported spending is not in the ground
Hold that number. First: who lost what?
Olken (2007). We come back in Part II to how he measured this.
What exactly was lost?
Author’s illustration. Only the top arrow moves money from one pocket to another. The rest is value that nobody recovers.
Today: three questions
- Does this make whole countries poorer?
- How does anyone know the money is missing?
- Who ends up paying the bribes?
The hard part is not finding a pattern. It is trusting one.
I. Does corruption make countries poorer?
One permit office, two stories
The “grease” story
- The legal process takes six months
- A bribe gets your permit tomorrow
The “sand” story
- Officials slow the queue so that people will pay
- Honest firms give up and go elsewhere
Grease: Leff (1964); Huntington (1968). Sand: Rose-Ackerman (1978); Shleifer and Vishny (1993).
A bribe can speed up one case while making the whole office worse.
The pattern everyone starts from
Mauro (1995): the first big test
- Analysts scored corruption in about 67 countries, 1980–83
- He lined those scores up against investment and against growth
- Growth was measured over 1960–85
- Nobody assigned countries their score by design
Mauro (1995), Section II. The scores came from Business International, a commercial country-risk service; a high score meant clean government.
What Mauro found
- Less corrupt countries invested more
- A big improvement went with investment about 3% of GDP higher
- Less corrupt countries also grew faster
- The investment result was the firmer of the two
Mauro (1995), Sections III.1 and III.3. “A large improvement” means one standard deviation — 2.51 points on his 0–10 scale, roughly the gap between Bangladesh and Uruguay in his data. Mauro himself reports the growth result as the weaker one.
Three stories, one pattern
| Story | What it says | An example |
|---|---|---|
| 1. Corruption makes countries poor | Investors avoid places where officials demand money | The factory gets built somewhere else |
| 2. Poverty makes corruption hard to control | Badly paid officials have more reason to charge | A clerk who cannot live on the salary |
| 3. Something else causes both | A weak state can neither collect taxes nor police itself | War, or courts that do not work |
Author’s illustration. Story 2 is what researchers call reverse causality; story 3 is a confounder.
The chart cannot tell these three apart. That is the whole problem.
And the ruler itself may be bent
- The score records what experts believe, not what happened
- Experts may simply assume that poor countries are corrupt
- Measured bribery and perceived corruption often disagree
Olken (2009) compares villagers’ perceptions with measured theft; Treisman (2007) compares what predicts perceived corruption with what predicts reported bribes.
We may be explaining income with somebody’s other judgement about income.
Exercise 1 · Which story?
In pairs — 5 minutes
Countries with more corruption are poorer. What could that mean?
- Pick the story you find most convincing: 1, 2, 3, or “all three”.
- Give one concrete example of your story.
- What would you have to see to change your mind?
Be ready to defend your choice against the other two.
So what?
- A pattern across countries cannot tell us which story is right
- Adding more control variables will not settle it
- So: measure corruption directly, in one place, against a comparison group
Next: what happens when somebody digs up the road to check.
II. Digging up the road
Olken (2007): 608 Indonesian villages
- A national village-grant programme, funded by the World Bank
- 608 villages in East and Central Java, 2003–2004
- Most projects: surfacing a dirt road with sand, rocks and gravel
- By lottery, some villages were told their project would be audited
- Audit chance moved from 4% to certain
Photo: Aryphrase, CC BY-SA 3.0, via Wikimedia Commons. Illustrative; not a project road from the study.
How you weigh a stolen road
%%{init: {"flowchart": {"useMaxWidth": true, "nodeSpacing": 120, "rankSpacing": 80, "padding": 48, "wrappingWidth": 700, "curve": "basis"}, "themeVariables": {"fontSize": "20px", "fontFamily": "Verdana, Geneva, sans-serif"}, "markdownAutoWrap": false, "width": 1150, "height": 620}}%%
flowchart LR
V["What the village reported"] --> D["Subtract"]
E["What the engineers found"] --> D
D --> M["24% missing"]
classDef box fill:#f9fafb,stroke:#4a7c6f,color:#1e293b,stroke-width:2.5px;
classDef hot fill:#b44527,stroke:#b44527,color:#f9fafb,stroke-width:2.5px;
class V,E,D box;
class M hot;
Author’s illustration of Olken (2007), pp. 202–203. The engineers’ figure came from digging core samples out of the road, pricing materials with local suppliers, and asking the workers what they were paid.
Audits cut the theft
What we learned, and what we did not
What we learned
- The comparison was made by a lottery, not by choice
- Announced audits cut missing spending by about a third
- The theft prevented was worth more than the audits cost
What we did not learn
- Audits did not end corruption: about a fifth still went missing
- Village accountability meetings had no average effect
- One programme, one country, one kind of road
Olken (2007), Sections V and VI. Theft prevented ≈ $468 per village against an audit cost of ≈ $335.
Bribes and firm growth in Uganda
- 243 manufacturing firms were asked, in confidence, what they paid in bribes
- Firms that paid more in bribes grew more slowly
- One extra point of bribes cost about three points of growth
- The bribe penalty looked larger than the tax penalty
Fisman and Svensson (2007), Ugandan Industrial Enterprise Survey; growth in sales, 1995–1997. Bribes and taxes are both measured as a share of the firm’s sales.
Why a bribe can hurt more than a tax
| A tax | A bribe |
|---|---|
| Known in advance | Secret and unpredictable |
| The same rule for everyone | Whatever the official thinks you can pay |
| Can be budgeted for | Can be demanded again next month |
| May come back as roads and schools | Buys no receipt and no protection |
Author’s summary of the argument in Shleifer and Vishny (1993), tested by Fisman and Svensson (2007).
This is why “corruption is just an extra tax” understates the damage.
So what?
- Measuring one thing twice turns a vague claim into a number
- A quarter of a road budget; three points of firm growth
- But those are averages, and averages hide people
Next: the same bribe is not the same burden.
III. Who pays?
Background: Ambrogio Lorenzetti, The Effects of Bad Government on the Countryside (detail), Palazzo Pubblico, Siena. Public domain, via Wikimedia Commons.
.JPG){kind=link}
Hunt (2007): asking households in Peru
- More than 18,000 Peruvian households, surveyed in 2002 and 2003
- Each was asked about its dealings with public officials
- And about any misfortune in the past year
- Outcome: did an official ask for, or take, a payment?
Photo: Deliaanicama, CC BY-SA 4.0, via Wikimedia Commons. Illustrative: Hunt’s data are national, not from this station.
Corruption finds you when you are down
Misfortune sends you to an office
- A robbery → the police
- An illness → the public hospital
- A death → the registry
- Job loss → the benefits office
Why a flat bribe is a regressive tax
Two things that are both true
- A bribe takes a bigger share of a poor household’s income
- Richer households deal with officials more often, and pay more in total
- So the total money burden is not concentrated on the poor
- But the poor have far fewer ways to avoid the public office
Hunt and Laszlo (2012), using household data from Peru and Uganda.
The same bribe is an inconvenience for one household and an emergency for another.
Exercise 2 · Find the hidden cost
In pairs — 5 minutes
Pick one case:
- A patient avoids the clinic because staff ask for money.
- A robbery victim never reports the crime.
- A small firm gives up on its permit.
Then answer: Who is harmed? What do they lose? Would that loss show up in government spending or in GDP?
The costs we still cannot price
- A crime not reported, because reporting costs a bribe
- A clinic avoided, a permit abandoned, a case never filed
- Talented people choosing rent-seeking over building things
- Citizens who conclude that the state is not theirs
Bribery after misfortune: Hunt (2007). Talent diverted into rent-seeking: Murphy, Shleifer and Vishny (1991), cited by Mauro (1995).
None of these show up in GDP. Each follows from something we have seen.
What we can say, and how sure we are
| What we can say | Where it comes from | How sure? |
|---|---|---|
| A quarter of a village road budget can vanish | Engineers dug up the roads | Very sure, for those villages |
| Announced audits cut the theft | Villages picked by lottery | Very sure, for those villages |
| Bribes slow firms more than taxes do | A firm survey in Uganda | Fairly sure |
| Bribery lands on people already in trouble | 18,000 Peruvian households | Fairly sure |
| Corruption makes whole countries poorer | Comparing countries with each other | Least sure |
Author’s summary of the assigned readings. “How sure” refers to the research design, not to the authors’ care.
Four costs, one lesson
- Money disappears
- Public services get worse
- People and firms stop using them
- The burden is heaviest where the alternatives are fewest
The bribe is the easiest cost to see — and the smallest part of the damage.
Next session: if the costs are this clear, why is corruption so hard to remove?
References
Fisman, R., & Svensson, J. (2007). Are corruption and taxation really harmful to growth? Firm level evidence. Journal of Development Economics, 83(1), 63–75. https://doi.org/10.1016/j.jdeveco.2005.09.009
Hunt, J. (2007). How corruption hits people when they are down. Journal of Development Economics, 84(2), 574–589. https://doi.org/10.1016/j.jdeveco.2007.02.003
Hunt, J., & Laszlo, S. (2012). Is bribery really regressive? Bribery’s costs, benefits, and mechanisms. World Development, 40(2), 355–372. https://doi.org/10.1016/j.worlddev.2011.07.015
Huntington, S. P. (1968). Political order in changing societies. Yale University Press.
Leff, N. H. (1964). Economic development through bureaucratic corruption. American Behavioral Scientist, 8(3), 8–14. https://doi.org/10.1177/000276426400800303
References (continued)
Mauro, P. (1995). Corruption and growth. The Quarterly Journal of Economics, 110(3), 681–712. https://doi.org/10.2307/2946696
Murphy, K. M., Shleifer, A., & Vishny, R. W. (1991). The allocation of talent: Implications for growth. The Quarterly Journal of Economics, 106(2), 503–530.
Olken, B. A. (2007). Monitoring corruption: Evidence from a field experiment in Indonesia. Journal of Political Economy, 115(2), 200–249. https://doi.org/10.1086/517935
Olken, B. A. (2009). Corruption perceptions vs. corruption reality. Journal of Public Economics, 93(7–8), 950–964. https://doi.org/10.1016/j.jpubeco.2009.03.001
Rose-Ackerman, S. (1978). Corruption: A study in political economy. Academic Press.
References (continued)
Shleifer, A., & Vishny, R. W. (1993). Corruption. The Quarterly Journal of Economics, 108(3), 599–617. https://doi.org/10.2307/2118402
Transparency International. (2025). Corruption Perceptions Index 2024. https://www.transparency.org/en/cpi/2024
Treisman, D. (2007). What have we learned about the causes of corruption from ten years of cross-national empirical research? Annual Review of Political Science, 10, 211–244. https://doi.org/10.1146/annurev.polisci.10.081205.095418
Popescu (TEC) Politics of Corruption Lecture 5: The Costs of Corruption