美國(guó)智庫(kù)發(fā)展經(jīng)驗(yàn)對(duì)中國(guó)特色新型智庫(kù)建設(shè)的啟示
——專訪華盛頓發(fā)展績(jī)效研究所高級(jí)研究員雷蒙德·J.斯特魯伊克博士

■ 記者：呂青 欒瑞英

美國(guó)智庫(kù)發(fā)展經(jīng)驗(yàn)對(duì)中國(guó)特色新型智庫(kù)建設(shè)的啟示
——專訪華盛頓發(fā)展績(jī)效研究所高級(jí)研究員雷蒙德·J.斯特魯伊克博士

■ 記者：呂青 欒瑞英

編者按：2016年12月17日，在由南京大學(xué)、光明日?qǐng)?bào)社主辦的2016中國(guó)智庫(kù)治理論壇召開(kāi)之際，我們有幸采訪了出席此次會(huì)議的雷蒙德?J.斯特魯伊克博士（Raymond J. Struyk），雙方就高端智庫(kù)建設(shè)、中美智庫(kù)區(qū)別、智庫(kù)獨(dú)立性、影響力、高質(zhì)量研究成果、智庫(kù)評(píng)價(jià)與排名等問(wèn)題進(jìn)行了短暫而愉快的交談，以期為中國(guó)新型智庫(kù)建設(shè)提供參考。

【訪談專家簡(jiǎn)介】

雷蒙德?J.斯特魯伊克（Raymond J. Struyk）是一位在發(fā)展中國(guó)家和轉(zhuǎn)型期國(guó)家研究方面經(jīng)驗(yàn)豐富的經(jīng)濟(jì)學(xué)家。1977年，他加入城市研究所，并于1981年開(kāi)創(chuàng)城市研究所的國(guó)際項(xiàng)目。1980年，他加入了卡特政府的住房和城市發(fā)展部，擔(dān)任負(fù)責(zé)研究和評(píng)估事務(wù)的副助理部長(zhǎng)。2007年至2012年，他擔(dān)任美國(guó)芝加哥大學(xué)國(guó)家民意研究中心高級(jí)研究員，致力于住房融資、項(xiàng)目評(píng)估以及智庫(kù)制度化發(fā)展的研究。他曾幫助匈牙利和俄羅斯建立智庫(kù)，并參與了十幾個(gè)關(guān)于管理、溝通和研究議題的具體工作。斯特魯伊克博士擁有圣路易斯華盛頓大學(xué)經(jīng)濟(jì)學(xué)博士學(xué)位，現(xiàn)為華盛頓發(fā)展績(jī)效研究所高級(jí)研究員，學(xué)術(shù)著作頗豐。

1 中美智庫(kù)的區(qū)別

問(wèn)：中美智庫(kù)之間有哪些差別？

答：中美智庫(kù)最大的區(qū)別是，美國(guó)智庫(kù)已經(jīng)用多年的時(shí)間來(lái)進(jìn)行管理實(shí)踐，也與政策界建立了良好的交流機(jī)制。對(duì)于中美智庫(kù)之間的差異這個(gè)問(wèn)題我感觸頗深。最近我在江蘇省考察了6家智庫(kù)，我以為這6家智庫(kù)都是非政府組織，他們的管理實(shí)踐豐富多樣。其中一家是大學(xué)智庫(kù)學(xué)院的所有教師同時(shí)也是智庫(kù)的工作人員。智庫(kù)管理者鼓勵(lì)他們寫(xiě)3000字左右的文章放在微信公眾平臺(tái)上，每周發(fā)布3~4篇，管理者對(duì)這種經(jīng)營(yíng)方式感到非常自豪，但這種方式所產(chǎn)生的影響力并不理想。此外，智庫(kù)管理者比較年輕，經(jīng)驗(yàn)不足，他們需要時(shí)間來(lái)學(xué)習(xí)如何管理一家智庫(kù)。當(dāng)然，也有正面例子。有兩家智庫(kù)組織結(jié)構(gòu)較為完善，并且與政府建立了良好的溝通渠道。其中一家智庫(kù)最初把團(tuán)隊(duì)建立在政府辦公室，但隨后把辦公地點(diǎn)搬至大學(xué)內(nèi)?，F(xiàn)在，工作人員在為一家智庫(kù)而不是一個(gè)政府機(jī)構(gòu)工作，他們擁有了相對(duì)自由的工作氛圍，并且能夠在傳播他們的研究成果中追求更廣泛的受眾。他們高度關(guān)注政策問(wèn)題，并希望將其研究成果與這些具體問(wèn)題對(duì)接。

它們像其他智庫(kù)一樣，面臨著如何避免資助機(jī)構(gòu)提出小而瑣碎的任務(wù)要求這一個(gè)普遍的問(wèn)題。我們大多數(shù)人知道：你是工作人員，你會(huì)被告知做某事，并且也沒(méi)有人想做。解決方案是把任務(wù)推給智庫(kù)。智庫(kù)避免資助組織不斷提出臨時(shí)請(qǐng)求的一個(gè)方法是，智庫(kù)制定一系列未來(lái)2~3年內(nèi)需要處理的優(yōu)先問(wèn)題以及實(shí)現(xiàn)這一目標(biāo)的戰(zhàn)略。資助者應(yīng)該審閱和同意該計(jì)劃。一旦戰(zhàn)略被正式采納，當(dāng)智庫(kù)從它的資助者那里收到過(guò)多數(shù)量的“意外任務(wù)”時(shí)，可以向其解釋它相關(guān)的機(jī)會(huì)成本，即如果要完成新任務(wù)，那么將拖延執(zhí)行戰(zhàn)略計(jì)劃。

政府機(jī)構(gòu)應(yīng)該制定一套智庫(kù)評(píng)估標(biāo)準(zhǔn)評(píng)價(jià)智庫(kù)的有效性和效率。顯然，這是一個(gè)需要現(xiàn)在完成的關(guān)鍵任務(wù)。此外，有一個(gè)論點(diǎn)是，一組明確的標(biāo)準(zhǔn)以及標(biāo)準(zhǔn)協(xié)議應(yīng)由智庫(kù)以及智庫(kù)所在省份的相關(guān)政府機(jī)構(gòu)聯(lián)合制定。這將鼓勵(lì)資助者之間知識(shí)共享，從而使協(xié)議有更大的可信度。

2 美國(guó)智庫(kù)建設(shè)經(jīng)驗(yàn)

問(wèn)：據(jù)我們所知，有些美國(guó)智庫(kù)接受政府委托的方案，會(huì)獲得政府資助。他們又是如何保持獨(dú)立性的？

答：這是一個(gè)好問(wèn)題。在卡特總統(tǒng)執(zhí)政期間，他任命我負(fù)責(zé)住房部的政策研究和計(jì)劃評(píng)估工作。在我為政府工作的這段時(shí)間里，我的辦公室每年（按現(xiàn)價(jià)計(jì)算）委托智庫(kù)和咨詢公司的計(jì)劃合同花費(fèi)約2億美元。智庫(kù)和咨詢公司需要通過(guò)競(jìng)標(biāo)拿到這些合同。

合同的條款對(duì)于智庫(kù)的獨(dú)立性至關(guān)重要。其中有兩個(gè)關(guān)鍵點(diǎn)，一是政府不能要求改變最終報(bào)告中的政策建議。當(dāng)然，監(jiān)督合同的官員有機(jī)會(huì)對(duì)報(bào)告提出意見(jiàn)。如果他們的意見(jiàn)是有價(jià)值的，智庫(kù)需要對(duì)報(bào)告做相應(yīng)修改。第二是智庫(kù)有公布最終報(bào)告的權(quán)力。即使政府不喜歡智庫(kù)提供的最終報(bào)告或沒(méi)有給承包商提任何意見(jiàn)，60天后，智庫(kù)仍有權(quán)發(fā)表報(bào)告。但政府也可能會(huì)反擊，之后不再選擇這家智庫(kù)接受政府的委托合同（在通常情況下極少發(fā)生）。

問(wèn)：美國(guó)智庫(kù)如何與政府進(jìn)行溝通呢？

答：有正式和非正式兩種溝通方式。就正式溝通而言，智庫(kù)可通過(guò)競(jìng)標(biāo)拿到政府的訂單，這就意味著智庫(kù)可以與政府進(jìn)行溝通了。另外，如果你一直研究A話題并成為這個(gè)方面的專家，那么政府可能會(huì)邀請(qǐng)你去向他人介紹你的發(fā)現(xiàn)或早期研究。這兩種情況都是官方正式行為。如果你足夠幸運(yùn)，政府還會(huì)將你推薦給國(guó)會(huì)相關(guān)人員。美國(guó)國(guó)會(huì)擁有大約12000名員工，525位官員，包括國(guó)會(huì)參議員和眾議員。其中有不少擁有博士學(xué)位的工作人員都接受過(guò)社會(huì)科學(xué)研究方面的培訓(xùn)。他們傾向于深入分析，例如閱讀報(bào)告并與作者交談，這樣智庫(kù)就有與國(guó)會(huì)議員溝通的機(jī)會(huì)了。而就非正式溝通而言，如果你是國(guó)會(huì)議員，我可以通過(guò)政府委派參加聽(tīng)證會(huì)而認(rèn)識(shí)你。之后你可以私下聯(lián)系我，講一下你在某個(gè)領(lǐng)域的新進(jìn)展，可以邀請(qǐng)我共進(jìn)午餐并探討一下相關(guān)問(wèn)題。類似的情況也發(fā)生在政府其他人身上。他們會(huì)帶一些人，你也會(huì)帶一些人，共同參加一個(gè)活動(dòng)，這樣兩個(gè)圈子的人可以逐漸了解和認(rèn)識(shí)。隨著時(shí)間的推移，你會(huì)建立起一個(gè)朋友圈。如果你在這行做了20年，那么你應(yīng)該會(huì)擁有豐富的人脈資源。那么你會(huì)很容易得到關(guān)于你這個(gè)行業(yè)的最新資訊。另一種非正式溝通是，國(guó)會(huì)通常會(huì)舉行大量的聽(tīng)證會(huì)，如果你跟國(guó)會(huì)相關(guān)人員很熟悉，他們可能會(huì)將你推薦給議員并邀請(qǐng)你參加聽(tīng)證會(huì)，長(zhǎng)此以往，你的簡(jiǎn)歷上就會(huì)有很多參加國(guó)會(huì)聽(tīng)證會(huì)的經(jīng)歷，那么人們看到這份簡(jiǎn)歷就會(huì)知道你是智庫(kù)領(lǐng)域的熱門(mén)人物。

問(wèn)：您的智庫(kù)是否擔(dān)心質(zhì)量不高的研究報(bào)告會(huì)影響后續(xù)合同的簽訂？

答：是的，會(huì)擔(dān)心。首先，我要說(shuō)優(yōu)秀的智庫(kù)會(huì)有非常嚴(yán)格的質(zhì)量管控。他們真的不希望報(bào)告質(zhì)量差。如果他們的報(bào)告得到政府的負(fù)面評(píng)價(jià)，他們會(huì)迅速做出回應(yīng)，消除這些評(píng)價(jià)。如果他們沒(méi)辦法消除這些負(fù)面評(píng)價(jià)，如果這是一份政府訂單，那么最糟糕的情況是政府拒絕接受報(bào)告，這意味著扣除應(yīng)支付給乙方費(fèi)用的20%~25%。當(dāng)然，政府通常設(shè)有評(píng)估委員會(huì)，他們負(fù)責(zé)評(píng)估提議和報(bào)告。但更糟糕的是，智庫(kù)和主要研究者的聲譽(yù)將受到嚴(yán)重破壞。如果團(tuán)體有不良聲譽(yù)，那么贏得下一個(gè)合同將是非常困難的。

問(wèn)：美國(guó)有新型高端智庫(kù)嗎？如何建設(shè)高端智庫(kù)？

答：我們對(duì)高端的定義跟您的可能有所不同。對(duì)我們來(lái)說(shuō)，高端意味著兩個(gè)方面：政策影響力和組織參與適量的活動(dòng)。首先，一個(gè)智庫(kù)可能在一個(gè)相對(duì)小的問(wèn)題上有卓越見(jiàn)解，而且頗具影響力，但這并不意味著這家智庫(kù)是高端智庫(kù)。另一點(diǎn)是，在我們的系統(tǒng)中要成為一個(gè)有影響力的智庫(kù)，必須能夠成功地籌集資金，因?yàn)檎惶峁┻@方面的預(yù)算支持。每個(gè)智庫(kù)都與其他智庫(kù)和咨詢公司去競(jìng)爭(zhēng)來(lái)自政府機(jī)構(gòu)和基金會(huì)的資金。大多數(shù)智庫(kù)都經(jīng)歷了擴(kuò)張、政策高影響力以及低潮的時(shí)期。因此，高端智庫(kù)不是永恒不變的稱號(hào)。智庫(kù)的政策影響力會(huì)隨著時(shí)間而變化?？赡芙?jīng)過(guò)若干年，一家富有影響力的智庫(kù)影響力逐漸減弱，因?yàn)樗麄兩瞄L(zhǎng)的政策議題不再被政府關(guān)注，因此即使他們推出了優(yōu)秀的成果，但依然沒(méi)有受眾。

我認(rèn)為你這個(gè)問(wèn)題里的重點(diǎn)應(yīng)該是“新”。我想到一個(gè)例子，華盛頓發(fā)展績(jī)效研究所（Results for Development Institute, 簡(jiǎn)稱R4D），僅僅成立六年。這個(gè)研究所的創(chuàng)始人曾經(jīng)就職于世界銀行,通過(guò)在發(fā)展中國(guó)家開(kāi)展示范項(xiàng)目來(lái)影響教育和健康等領(lǐng)域的新政策。這家智庫(kù)不是與政府直接合作，而是主要與國(guó)際組織、基金會(huì)、美國(guó)駐發(fā)展中國(guó)家辦事處合作。目前，這家智庫(kù)在很多國(guó)家都具有一定的影響力。

建設(shè)高端智庫(kù)，首先要明確智庫(kù)需要不斷創(chuàng)新，與時(shí)俱進(jìn)。幾乎所有智庫(kù)的創(chuàng)立初衷都是因?yàn)樗麄兊膭?chuàng)始人認(rèn)為某個(gè)問(wèn)題沒(méi)有得到足夠的關(guān)注。在歐洲的智庫(kù)亦是如此。他們最初幾乎只專注于某一政策領(lǐng)域，逐漸形成一定的影響力。然后，開(kāi)始擴(kuò)展到其他政策領(lǐng)域。設(shè)立的新議題或新項(xiàng)目會(huì)接近過(guò)去研究的核心議題。隨后他們?cè)O(shè)立更多的議題，雇傭更多的工作人員。智庫(kù)在強(qiáng)大的管理者的領(lǐng)導(dǎo)下，還需要開(kāi)展7方面的工作，具體如下。

（1）定義智庫(kù)的使命。圍繞這個(gè)使命，雇傭工作人員，同時(shí)，這個(gè)使命可以為開(kāi)展具體項(xiàng)目提供重要指導(dǎo)。

（2）保持獨(dú)立性。只有真正獨(dú)立的智庫(kù)才會(huì)做出強(qiáng)有力的客觀分析，他們的結(jié)論和建議是基于證據(jù)而非政治或個(gè)人意見(jiàn)。確保提供持續(xù)的客觀公正的政策建議的重要舉措要從智庫(kù)的董事會(huì)開(kāi)始。董事會(huì)可以在智庫(kù)簽訂合同或項(xiàng)目之前審查所有大額贈(zèng)款，堅(jiān)持更改或拒絕不符合此標(biāo)準(zhǔn)的相關(guān)條款。

（3）招聘優(yōu)秀的政策分析師，以確保研究的專業(yè)性。這些優(yōu)秀的分析師一直對(duì)他們的研究和政策制定策略采取嚴(yán)格的循證分析方法。

（4）嚴(yán)格的質(zhì)量控制至關(guān)重要。以防止不直接基于分析結(jié)果的結(jié)論被公布或以其他方式被提倡。如果智庫(kù)工作的早期階段出現(xiàn)質(zhì)量問(wèn)題，很難恢復(fù)聲譽(yù)。

（5）積極參加各種高層論壇和會(huì)議。主動(dòng)與潛在的贊助商和政策制定者進(jìn)行交流，描述相關(guān)倡議，并就可能研究的新政策問(wèn)題征求建議。在這個(gè)階段不要直接進(jìn)行籌款。這些會(huì)議是進(jìn)入政策決策層的關(guān)鍵。

（6）盡早建立官方網(wǎng)站，并不斷豐富形式，完善內(nèi)容。如果你不定義自己，展現(xiàn)自己，別人的錯(cuò)誤描述可能會(huì)影響智庫(kù)的聲譽(yù)。必須通過(guò)網(wǎng)絡(luò)平臺(tái)展示自己的獨(dú)立性和影響力。

（7）定期與所有工作人員溝通，不斷完善智庫(kù)管理機(jī)制。

問(wèn)：您對(duì)美國(guó)智庫(kù)發(fā)展態(tài)勢(shì)有何看法？

答：據(jù)我粗略估計(jì)，美國(guó)現(xiàn)在有75%的智庫(kù)是自主型和獨(dú)立型。15%~20%的智庫(kù)是半獨(dú)立型。出現(xiàn)半獨(dú)立型智庫(kù)的原因是一些富有的資助者和團(tuán)體支持認(rèn)同他們的政策哲學(xué)或與他們的政治方向一致的智庫(kù)。他們可能為智庫(kù)提供非常大的資助。這些智庫(kù)從發(fā)布基于證據(jù)的政策報(bào)告轉(zhuǎn)向了銷(xiāo)售預(yù)先確定的立場(chǎng)的游說(shuō)產(chǎn)品。鑒于這種情況，一些基金會(huì)已經(jīng)采取了行動(dòng)，迫使所有智庫(kù)保證資金來(lái)源透明化。這個(gè)想法是，如果研究的發(fā)起人是對(duì)結(jié)果有直接利益關(guān)系的組織或個(gè)人，那么閱讀報(bào)告或者考慮報(bào)告中的政策建議的人可能會(huì)懷疑這份報(bào)告的公正性。例如，一家石油公司贊助一個(gè)與石油污染相關(guān)的項(xiàng)目。幾個(gè)基金會(huì)創(chuàng)立了一個(gè)網(wǎng)站叫Transparify.org，他們關(guān)注智庫(kù)網(wǎng)站公布的經(jīng)費(fèi)來(lái)源。

我認(rèn)為美國(guó)智庫(kù)的政治化傾向可以被控制。這將需要基金會(huì)、其他民間社會(huì)組織以及智庫(kù)本身的不懈努力。如果政治化的影響繼續(xù)下去，智庫(kù)可能會(huì)失去在過(guò)去60年來(lái)所享有的在政策制定過(guò)程中的特殊地位。

3 智庫(kù)評(píng)價(jià)對(duì)智庫(kù)發(fā)展的作用

問(wèn)：您認(rèn)為智庫(kù)的評(píng)價(jià)與排名，對(duì)智庫(kù)發(fā)展有多大作用？

答：很難做這樣的比較。但這種比較可以提供有價(jià)值的信息。最大的智庫(kù)評(píng)價(jià)項(xiàng)目是由賓夕法尼亞大學(xué)詹姆斯·麥甘（James G. McGann）領(lǐng)導(dǎo)的。他的研究方法是保密的。因此，有些人不會(huì)非常重視排名。而比較通用的方法是設(shè)定普適的指標(biāo)，例如智庫(kù)在過(guò)去兩年在同行評(píng)議期刊上發(fā)表的文章數(shù)量，參加國(guó)會(huì)聽(tīng)證的次數(shù)，參加電視訪談的人數(shù)，報(bào)紙引用次數(shù)。你也可以問(wèn)相關(guān)智庫(kù)，因?yàn)榇蟛糠种菐?kù)保留了他們接到政府約談的記錄。而智庫(kù)的產(chǎn)出和活動(dòng)可以作為評(píng)估智庫(kù)影響力的間接指標(biāo)，但確定智庫(kù)是否真正影響決策是非常困難的?？傊?，我認(rèn)為人們不應(yīng)該過(guò)分看重這些排名。

問(wèn)：您如何評(píng)價(jià)由李剛教授主導(dǎo)的CTTI中國(guó)智庫(kù)索引？

答：我認(rèn)為這個(gè)索引很有用，也會(huì)成為真正的資源。如果想要找到某一個(gè)特殊話題的專家，那么可以查閱出版物，然后找到專家。這個(gè)索引可以促進(jìn)智庫(kù)研究工作的開(kāi)展。同時(shí)，它可能也會(huì)引發(fā)智庫(kù)專家的跳槽現(xiàn)象。因此，這是一把雙刃劍。我不知道他們是否有國(guó)際期刊評(píng)價(jià)這種黃金評(píng)價(jià)指標(biāo)。如果談研究成果的質(zhì)量，那么發(fā)表國(guó)際期刊論文無(wú)疑是高質(zhì)量的體現(xiàn)。

簡(jiǎn)而言之，這個(gè)數(shù)據(jù)庫(kù)具有成為極有價(jià)值資源的潛力。

附采訪稿英文原文

1 The difference between Sino and USA think tanks

Question (Q): What is the biggest difference between Sino and USA think tanks?

Answer (A): The major difference, of course, is that USA think tanks have had many more years to develop their management practices and build bridges to the policy communities compared with their Chinese counterparts. I think a better question here is one about the differences among Chinese think tanks.

This week I met with six think tanks all in Jiangsu Province. I think they are all public NGOs, and no social NGOs were included. These sessions were for wide-ranging discussions of management practices, not me making presentations. Even with this tiny sample of six, a wide range of diversity in management practices was evident. I did not read any reports, on how they are organized, or on how they are setting their agenda or similar questions, but the discussions gave me some basic information.

The difference in some practices among these six is striking. Let me give a couple of examples without giving the names of the think tanks. At one university, all faculty are on the think tank’s staff automatically. The staff is encouraged to write essays over around 3,000 words that are put on the WeChat public platform—3-4 essays a week. And the managers are very proud of this productivity. As described to me there is no particular focus, no follow-up. No quality control was outlined, although it may exist. This approach is unlikely to result in much policy change. The people managing the organization are pretty young. They will learn and improve operations, but this will take time.,

On the other hand, I met with two think tanks that appear to have very good structures. They evidently have close relationships with the “right” government officials. One of these was initially staffed, by taking the team in a government office and moving it to the university to be the think tank’s core staff. Now the staff works for a think tank rather than a government agency and seems to have greater freedom in deciding the topics on which they will work and may be able to pursue a broader range of audiences in disseminating their work. They clearly focus on policy questions, and want to bring their research results to bear on these specific issues. They are pursuing“evidence-based policy making.” The individuals with whom I met seem to be really enjoying working in the new environment.

Part of the conversation was how they can protect their institute from getting endless requests to do small tasks for the ministry that provides its funds. This is a familiar arrangement that most of us know: You are the staff in the ministry, and you are told to do something and no one wants to do it. The solution is to push the task to the think tank. One approach for the think tank to control the fl ow of ad hoc requests from the funding organizations is to develop a set of priority issues that it will address over the next 2-3 years and a strategy for accomplishing this. The funder should, naturally, be consulted and agree with the program. Once the strategy is formally adopted,the think tank can, when it receives an excessive number of “surprise tasks” to do from its funder, explain the opportunity cost associated with doing it—if the new task is done, then there will be delays in carrying out the strategy.

The reality is that is too early to judge the productivity of most of these think tanks. One or two of the six I met with had been founded only in 2016. I do not know if any government agency is developing a set of criteria that can be applied in another year or two to rate their effectiveness and efficiency. Obviously this is a key task that needs to be done just now. Moreover, there is an argument that a standard set of criteria and a standard protocol for applying the criteria should be developed jointly by a group of key agencies from a group of provinces working with think tank leaders. This would encourage knowledge sharing among funders, give the protocols to develop greater credibility, and improve consistency across think tanks in their ratings.

2 The construction experience of American think tanks

Q: As far as we know, there are some think tanks to receive programs entrusted from the government and take money from the government. How can they keep independence?

A: Good question. It is one about which I know something. When President Carter was in the White House, he appointed me to be responsible for the policy research and program evaluations at the housing ministry. I worked for the government in these years. My office commissioned about 200 million dollars a year (in current prices) on contracts to think tanks and consulting firms for research and evaluations of the ministry’s programs. All awards were done through competitions.

The terms of a contract are critical for think tanks’ independence. Two terms were especially important. One fundamental term was that the government could not demand changes to policy recommendations in final reports. Of course, the officials overseeing the contracts had the opportunity to provide comments on reports. And if their comment had merit, the contractor would want to make the suggested change; that is fine, but they could not demand them except on narrow technical points. The second key term was the contractor’s right to publish the final report. Even if the government did not like it or did not give the contractor comments, after 60 days, the contractor still had the right to publish the report. The government could not block it. If you are a contractor, you do not want to do this too often, because presumably the government will fight back, not choosing you again for a new contract (even though this is officially not to happen).

Q: How does a think tank exchange information with the government?

A: There are two tracks—formal and informal. On the formal track, there are things like a competition for a new research or evaluation contract or formal requests from the government agency for a think tank to consult with it on a specific policy issue. There may also be a request from a staff member of the U.S. Congress to meet with you about research you are doing that is relevant to policy issues for which he has some responsibility. The Congress has about 12,000 staff members (there are 525 congressmen and senators in total). There are many staff with a Ph.D who are trained in social science research. They tend to want to dig fairly deeply into the analysis by reading the report and talking with the authors and so on.

On the informal track, analysts develop relationships with officials; these exchanges serve both the analyst and officials very well in broadening their thinking on specific policy issues. If you work consistently on subject A, then you are likely to be viewed as one of the experts in that subject. Officials will ask you to talk to them and others about what you are finding or what your early research shows on this topic. These are accepted as routine exchanges. If you are a staff member to the Congress, I get to know you by being brought over by the ministry. After that I can follow up, perhaps giving you a call and saying, “You know, we have a little more work on this topic, how about we discuss this over lunch?” The same process works with the people in the ministries. Often they will bring somebody with them, and you bring somebody with you. If you work at it steadily, you build upyour contacts, and you can be quite effective in the policy process. If you have been in the game for 20 years, you will have a lot of contacts and you can find out what is currently going on in your field and what topics will be “hot”in the months ahead.

Especially after I worked in the ministry, in the next ten years I had a number of opportunities that emerged from the informal contacts. The Congress holds a lot of hearings where Members of the Congress discuss policy issues with experts. You may have seen these hearings on TV. If you get to know the congressional staff well, they will recommend you to the committee chairmen to be an expert witness. When someone sees on your resume a list of congressional hearings where you have given testimony, he understands you are well-connected in the policy process.

Q: Does your think tank worry about that unsatisfactory research reports will affect follow-up order receiving?

A: Yes. This is a major concern for all think tanks in the USA. First of all, a good think tank has very strong quality control. It really does not want to turn in a bad report. When the researchers receive critical comments from reviewers, they really have to respond to them carefully and fully.

But a think tank can submit a poor quality reports because of a weak job on quality control, which does happen from time to time despite the fact that official policies that should prevent it. In the worst cases, if it is a government contract, the government will not accept the work, and the contractor will not be paid the last 20 percent or 25 percent of the contract amount. But even worse, the think tank’s and the primary researchers’reputations will be badly damaged within the ministry. If your reputation with the group in the ministry that reviews proposals is bad, then winning the next contract will be very hard.

Q: Does the USA have new high level think-tanks favored by the government? What kind of think tanks should be constructed in China today?

A: We do not have the concept of the “high level think tank”. High-level for us would mean a combination of two things—policy influential and a reasonable volume of activities. A think tank may be quite effective on a single and very narrow topic, but that alone does not make you a leading think tank. That is one point. The other point is to be an influential think tank in our system, the think tank has to be successful in raising money, because the government does not provide budget support. Each think tank competes against other think tanks and consulting firms for funds from government agencies and foundations. Hence, most think tanks experience times of expansion and policy influence and times of less success.

The accent in your question was on a “new” think tank, I think. I can think of one. It is only six years old. It is called Results for Development Institute (R4D). It was started by a fellow who just left the World Bank as a vice president and was primarily interested in implementing demonstration projects in developing countries as a tool for supporting new policies, especially in the health and education sectors. But he obtained support, not from our government, but from international organizations and foundations primarily, as well as, later, one U.S. government agency that works in developing countries. This think tank’s success in having genuine impact in a number of countries is an astonishing story. Eight people from Jiangsu Province came to Washington in September, where they met with R4D’s founding president and were impressed by R4D’s approach.

A key point is that think tanks’ effectiveness varies over time. One does not know if an organization will be more or less influential over time. It is possible that the policy issues where some think tanks are strongest are of little interest to the government; so they may be doing good work, but there is no audience. Hence, as for influence, it is really quite variable across think tanks and over time for the same think tank.

Think tanks keep getting created. Nearly all start because their founders believe that a certain issue is not being given enough attention. I can tell you that this is the case in the US. It is also the same in Europe and Eastern Europe. I have worked in these places and have some first-hand knowledge. They focus initially almostexclusively on their core policy area and, with luck, they have some impact. Then they begin spreading out into other policy areas. Typically the new areas are close to the core area so that current staff can work on both. As they take on more topics, they hire more staff.

Besides attracting a strong executive director, there are several other critical tasks in starting a think tank.

（1）Take the time and energy to define the think tank’s mission carefully—even if all the work done in the early years may not be perfectly consistent with it. It provides essential guidance for selecting work, hiring staff, project execution, and action in the policy arena.

（2）Visit potential sponsors and policymakers to describe your initiative and solicit advice on emerging policy issues you might study and on management tips. Do not ask for funds at this stage. These meetings are key to put you into the policy community consciousness.

（3）Focus on a few areas of high priority for the policy community to attract attention. Usually one needs to deliver a critical mass of advice in an area to be taken seriously, and can start research on other areas in the meantime.

（4）Strong quality control is absolutely critical from the very start. It is very hard for a think tank to recover respect if it has early problems with the quality of its work.

（5）Get a communication operation up early. If you do not define yourself, others will and their descriptions are likely to be inaccurate. It is essential to be clear about the relationship with the sponsor. You will need to demonstrate your independence from sponsors.

（6）Attract a few very good policy analysts to ensure professionalism in the organization’s work and to have an image of competence. Consultants are the second best because they can be hired by anyone and therefore, reports they prepare for you only strengthen your reputation to a limited degree.

（7）Governance:①establish a tradition of group decision making (not senior staff voting, rather consultation); ② communicate regularly with all staff—lots of options on how to organize this.

Q: What do you think about the development trend of American think tanks?

A: For the last ten years or so, this autonomous and independent group has been very dominant..

My rough estimate is that today still 75% of think tanks belong to this category. But15%-20% are in another category, quasi independent. This category has arisen because some wealthy donors and groups support think tanks that broadly share their policy philosophy or the political direction they favor. They might offer a think tank a very large grant for a new center on the environment or another topic of interest to the donor. If it is a conservative donor, the expectation is that the new center will fight against any more regulations that restrict what businesses can do. The think tank is expected to produce evidence that the costs of more restrictions on pollution will be too expensive for the economy. These think tank move away from evidence-based policy formulation to one where they are, in effect, selling a pre-identified position. Think tanks of this type have both conservative and liberal orientations.

Given this situation, civil society is thinking about what can be done to try to slow down this movement away from evidence-based proposals. Some key actors are outside the think tank community. Foundations in particular have already taken some action, particularly in the area of pressing all think tanks to be transparent about the sources of their funds, both in general terms and for each project. The idea is that someone reading a report or considering policy recommendations in a report may be skeptical about them if the sponsor of the research is an organization or individual that has a direct interest in the outcome—for example, an oil company sponsoring an analysis of the pollution associated with drilling for oil. Several foundations found a new group called Transparify. org, and they look at how much information is on think tank websites about where their money comes from.

There are also critical actions that can be taken to help ensure a continuation of evidence-based policyrecommendations. This starts with the board of directors whose main task is to make sure the think tank’s behavior remains consistent with its mission, which nearly always includes taking an independent, objective approach to its analysis. One important task the board can perform is to review all large grants or contracts before they are signed to be certain that their goals and terms are consistent with the organization’s mission and insist on changing or rejecting those that do not meet this standard. The second key area is hiring and rewarding analysts who consistently take a rigorous evidence-based analytic approach to their research and policy development. A strong quality control system must be in place to prevent conclusions that are not based squarely on analytic results from being published or otherwise advocated.

My sense is that the drift among U.S. think tanks towards politicization can be arrested. But this is going to take a strong effort by foundations, other civil society organizations, as well as think tanks themselves. If the drift to politicization continues, think tanks will likely lose the privileged position in the policy process they have enjoyed over the past 60 years.

3 The effect of think tank appraisal on the think tank development

Q: Is it beneficial to the development of think tanks to appraise think tanks through comparisons that result in rankings?

A: It is hard to do such comparisons well. But such comparisons can provide valuable information that can be especially useful for those outside the think tank community. The biggest think tank ranking machine is this one at the University of Pennsylvania. The methodology employed is secret. As a consequence, many people do not take the rankings very seriously. On the other hand, all think tanks that receive high ratings tout them.

An alternative to the secret methodology is to restrict the data employed to widely accepted indicators, such as how many articles a think tank published in the past two years in peer-reviewed journals, how many people testified before the parliament, how many TV appearances did staff make, how many mentions of the think tank appeared in newspapers. One could also learn from think tanks’ information they gather for internal use, such as how many appointments they had with ministry officials in the reference time period. Outputs and activities are indirect indicators of influence and so are only broadly effective in measuring success in the policy arena. But identifying who really influences something is extremely difficult.

In sum, I think one needs to look at these rankings as being only a very general indicator of success.

Q: What do you think of the CTTI database led by Professor Li Gang?

A: He gave me a tour of the new database for 30 minutes, but I do not claim to really understand it. It is useful and it can be a real resource. He pointed out, for example, if you are looking for someone who is expert in a particular topic, you can search publications to find the qualified people at think tanks. It has the potential to facilitate dramatically a lot of work involving multiple think tanks.

It may also result in strong analysts being hired away from their “home” think tank. You can look at such a transaction being positive or negative, depending on whether you are the winner or the loser in a particular case. The data base contains entries for individual analysts for many conference presentations and reports. I do not know the extent to which Chinese think tank analysts have been publishing in international refereed journals, which is the gold standard for quality publications. I am on the board of several journals and for these journals, at least, the incidence of authors with a Chinese name has increased very sharply in recent years..

In short, the data base holds the potential to be an extremely valuable resource. Of course, it will only achieve its potential if think tanks provide information on themselves to CTTI.

2017-01-16

2017-01-18 本文責(zé)任編輯：唐果媛

呂青（ORCID: 0000-0003-3101-1733），《智庫(kù)理論與實(shí)踐》編輯部主任，副研究館員，E-mail: luq@mail.las. ac.cn。欒瑞英（ORCID: 0000-0002-4721-9128），中國(guó)科學(xué)院文獻(xiàn)情報(bào)中心博士研究生，E-mail: luanruiying@mail.las.ac.cn。